smoido/claude-mode
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Compare commits

base: smoido/claude-mode:da50a3c7e5de8364a90abb09c183148a28a70f63
Branches Tags
smoido/claude-mode:master
...
compare: smoido/claude-mode:7dfc428d63e3764335c5f8645f09c48259bfdf5f
Branches Tags
smoido/claude-mode:master
10 Commits
da50a3c7e5 ... 7dfc428d63
Author SHA1 Message Date
smoidoandClaude Code 7dfc428d63 Install straight from the repo with one piped command 
Piped in via irm | iex there is no checkout next to the script, so the
installer now notices that and fetches the repository archive to %TEMP%
itself. README documents the one-liner, the download-and-run variant for
switches, and the clone path.

Co-Authored-By: Claude Code <noreply@anthropic.com>
 2026-09-01 21:43:57 +03:00
smoido 4c36f4d00b Notice broken sessions in the bar, and offer to repair them there 
Nothing tells you a session is unresumable until you try to resume it, and by
then you have usually forgotten which one it was. The widget now scans every
project on a timer and whenever the panel opens, marks its icon when something
needs fixing, and lists the affected sessions with a repair button that says
what it will drop and what it will keep before doing anything.

The scan had to get roughly eighty times cheaper first. Classifying a transcript
needs two facts - what its last message id is, and whether any Anthropic id
exists at all - and the first settles the common case alone. Reading the tail of
each file and only opening the whole thing when the tail already looks wrong
takes the sweep from ~5s to ~60ms across 59 transcripts, which is the difference
between something that can sit on a timer and something that cannot. `--all`
uses the same path, and `--json` exposes it.

The badge is a dot beside the mark rather than a recolouring of it: this widget's
job is to report which provider is active, and tinting it red to mean something
else entirely would be a lie about that.

Verified by planting a genuinely corrupted transcript, watching the scan find it
and the dot appear, then removing it and watching both clear.
 2026-08-31 02:28:24 +03:00
smoido 43cef6bf9e Scan every project for broken transcripts, and stop crying wolf 
repair-session only ever listed the project you were standing in, which is a
poor place to start from when the thing you cannot resume is a session whose
project you no longer remember. --all drops the working-directory scoping and
checks all of them. (A named session id was already looked up across every
project; only the listing was scoped.)

The first version of that scan reported 16 of 59 transcripts as unresumable,
which was true in the narrowest sense and useless in every other. Checking them:
5 had never received an assistant reply at all, and 10 had run start to finish
on a gateway, so every id in them is that provider's by design. Those resume
perfectly well under the provider they were born on, have nothing to truncate
back to, and are only a problem if you try to resume them as Anthropic. Calling
either of them damage buries the one case that is.

So the verdict now carries a kind - healthy, repairable, gateway-native,
synthetic-only, no-messages - and only `repairable` is reported or acted on: a
transcript with a genuine msg_ message and junk after it. Across the same 59 it
now correctly reports nothing to repair, while still classifying the real
corrupted transcript from the incident as repairable with 108 lines to drop.

A gateway-native session asked about directly now says what it is and that there
is nothing to fix, rather than failing with "no Anthropic-issued message".
 2026-08-31 02:22:01 +03:00
smoido a14418b5d9 Keep the turns a repair cuts, and hand them back to the session 
Truncating the transcript is the mechanical fix, but the turns being cut are the
work itself. Losing the conversation that produced a morning's changes is most
of the damage, and a session that resumes with a hole in its memory is barely
resumed - it has no idea what it just did or what it was asked.

So --apply now writes the dropped turns out as <session>.recovered-<stamp>.md,
a readable record of what was asked, what was answered and what was run, and
appends the same text back to the truncated transcript as a single note. Tool
results are deliberately excluded: they are most of a transcript by volume and
the least useful part of a summary.

The note is a user entry marked isMeta - the marker Claude Code already uses for
its own local-command caveats, meaning context rather than something to answer -
and carries no message.id, so it cannot recreate the previous_message_id
condition being repaired. Verified on the real corrupted transcript: 1921 lines
in, 1813 kept plus the note, chained to the last good assistant uuid, no id on
the message, and the file reads back as healthy. The generated digest recovers
the actual instruction that was lost in the incident, which is the thing that
made this worth doing.

--no-reinject writes the markdown but leaves the session alone.
 2026-08-31 02:14:42 +03:00
smoido a4afa55580 Ask before switching while sessions are live, and recover the ones already broken 
Acts on docs/incident-mode-switch-corrupts-live-sessions.md, which is added here
as the record of why.

The report identifies a consequence that was not modelled. A failed call is
recoverable; a *successful* one may not be. If a running session takes even one
completion from the provider being switched to - which happens when that mode
matches the base URL it already had cached - that provider's message-id format
lands in its transcript. OpenRouter issues `gen-<epoch>-<rand>` where Anthropic
issues `msg_...`, and native Anthropic then refuses to resume the session at
all, with a 400 naming previous_message_id. The only way back is to truncate the
transcript, losing every turn after the cut. That happened here, and was fixed
by hand.

Two changes follow.

Sessions are now settled before the write, not reported after it. A switch with
anything running stops, names the sessions, explains what is about to happen to
them, and offers restart (the only answer that ends with everything on the mode
the bar now claims), close, proceed anyway, or abort - defaulting to abort.
Non-interactively it refuses outright unless given --yes. The old after-the-fact
reporter is deleted rather than left as a second, contradictory account.

`claude-mode repair-session` replaces the hand surgery: it finds a project's
transcripts, reports which are resumable, and on --apply backs the file up and
truncates to the last Anthropic-issued message. Verified against the real
corrupted transcript from the incident - it reproduces the manual cut exactly,
1921 lines to 1813, dropping the two `gen-` completions and the error
placeholders after them, leaving a transcript that ends on a genuine msg_ id. It
refuses a transcript written to in the last 90 seconds, since that one belongs
to a session still running.

The panel passes --yes, having already asked in its own card, and that card now
names the transcript risk rather than only the inconvenient one.

Requirement 4 of the report - documenting the mechanism - landed in 9c301e1;
the README now carries the unrecoverable half as well.
 2026-08-31 02:05:37 +03:00
smoido 9ff9b82c25 One preset per mode, and a first-run setup that fills it in 
The openrouter default moves its hot tiers: opus to z-ai/glm-5.3-flash and
sonnet to deepseek/deepseek-v4-flash-0731. haiku and fable are unchanged.

`cheap` and `lmstudio-qwen` are gone, leaving exactly one preset per mode so
`claude-mode <mode>` is never ambiguous and there is no menu to read before the
thing you asked for happens. The surviving lmstudio preset keeps the Qwen3.6
model rather than KAT-Coder: the two differed mainly in that KAT's chat template
carries the message-order assertion this README already warns about, so between
two presets that had to become one, the one that is known to work won.

More presets are still a `preset new` away; the shipped set is a starting point,
not a ceiling.

Which is the other half of this. A shipped preset was never a working
configuration - OpenRouter and Z.AI have no key stored, and lmstudio's model ids
were whatever happened to be installed on the machine this was packaged on. That
was left for the user to discover through a failure. Now the shipped presets
carry `configured: false`, preflight blocks on it, and `claude-mode setup <mode>`
walks through what is actually needed: key, server URL and auth for LM Studio,
then models chosen from the provider's own catalogue rather than typed from
memory. A switch that trips this in a terminal offers to run setup there and
then instead of printing a command to type next.

Absent means configured, deliberately: presets that predate this and any built
by hand with `preset new` do not suddenly start demanding a wizard.

The panel gets a "Set up <mode>…" button that hands the whole flow to a terminal,
since a bar popup can host neither a hidden key prompt nor a filter-select list.

Two bugs found while testing it, both real:

ask_value printed its prompt to stdout while being called inside $( ), so the
prompt text came back glued to the front of the answer and set-url rejected the
result. Moved to stderr, which is why warn and err already go there.

lms_catalogue never sent the API key. On a server with authentication switched
on - the case just added support for - /api/v0/models answers 401 like anything
else, so the catalogue came back empty and every caller silently concluded the
server had no models installed. It now sends the preset's credential, as do the
three other call sites that read it.
 2026-08-31 01:48:17 +03:00
smoido 9c301e10a1 Correct what a switch actually does to a running session 
Every one of these said a switch leaves running sessions alone - that they keep
talking to the old provider until restarted. That is wrong, and reassuring in
exactly the wrong direction: a switch breaks them.

The static half of the config does behave as described. Base URL, model ids and
the env block are read once at startup and a running session keeps what it
started with.

The credential is not. It comes from running apiKeyHelper, which Claude Code
re-invokes on a timer - CLAUDE_CODE_API_KEY_HELPER_TTL_MS is present in the
2.1.251 binary - and claude-key-helper.sh answers for whatever state.json says at
that moment. So the switch reaches into a live session through the one part that
was never cached. Into anthropic, the helper returns nothing by design and the
next refresh comes back with no credential; into another provider, it hands over
the new key while the session is still pointed at the old base URL, which
rejects it. Either way the session starts failing calls whenever the TTL happens
to expire, mid-turn as easily as between turns.

One case does survive, and is now stated rather than glossed: switching between
two presets of the same provider that share a keyRef keeps the same key and the
same endpoint, so the session carries on with the model ids it started with.

The claim in "Restarting sessions" predates this work - it came in with the
Windows build - but it is wrong for the same reason and is corrected too.
 2026-08-30 21:41:09 +03:00
smoido 22cadaebc5 Wrap the panel's button rows instead of running past the card edge 
The session-confirmation card lays out four buttons, which come to roughly
440px of content in a card 332px wide (360 less 14px of padding each side). A
Row has no notion of how wide its parent is and no wrap, so it put them in one
line and the last two ran off the right edge of the popup.

Flow bound to the parent width instead, so the buttons take a second line when
they need one. Applied to every button row rather than just the one that
overflowed - the failure card can show four at once for the same reason, and the
server form's rows would do it at a narrower card or a larger font.

Labels are left as they are: "Switch and restart" says what the button does to
the sessions listed above it, and two tidy rows read better than four terse ones
crammed into a single line.
 2026-08-30 21:32:31 +03:00
smoido 35690bcd8d Move GLM 5.2 to 5.3 
GLM 5.3 supersedes 5.2, so every reference to the older id follows it: the zai
preset's opus, sonnet and fable tiers, and the fable tier of `default`, which
routes to the same model through OpenRouter.

Also the hardcoded Z.AI entries in both builds' model pickers, which are the one
place a model id is not read from a preset - leaving those would have gone on
offering a superseded model from the menu while the presets used the new one.

haiku stays on glm-4.7: that is the fast/cheap tier, not a 5.x.
 2026-08-30 21:28:45 +03:00
smoido f68ca08009 Omarchy bar widget: show the active provider, and switch from it 
A Quickshell plugin for the Omarchy shell (v4's bar is Quickshell, not Waybar).
The icon is the mode; clicking it opens a panel that switches.

State comes from watching health.json rather than polling the CLI, so the widget
costs nothing while idle and a switch made in a terminal shows up in the bar on
its own. state.json is the fallback, because it exists from install onward where
health.json only appears once claude-mode has run.

The marks are the providers' real logos, drawn as vector paths through
QtQuick.Shapes rather than set as font glyphs - three of the four have no Nerd
Font pictograph at all, and paths take the bar's foreground colour and follow
the theme. Claude, OpenRouter and LM Studio from simple-icons, Z.AI from
lobe-icons; trademarks belong to their owners. Zhipu's mark was rejected for
Z.AI: it is a dense lattice that is unreadable below ~40px.

Rendering them at 13px took three fixes. A layer rasterises the Shape at its own
size and then scales the texture, so a 24px buffer minified to 13 resampled two
pixels into one and the Claude burst lost rays; without the layer the scale is a
transform on the geometry and rasterisation happens once, at final resolution.
CurveRenderer replaces the tessellating default. And the size is forced odd,
because a radially symmetric mark puts its vertical and horizontal arms on the
centre line, which is a pixel centre at odd sizes and the seam between two
pixels at even ones, where each arm splits its coverage and greys out. Measured
against a cairo render at the same size, the result is now identical.

Sizing is measured rather than guessed. Every stock glyph in this bar paints
11px of ink; the marks fill their box instead of carrying a font's padding, so
the box is the smaller number. Each mark also carries an optical scale from two
measurements of a 200px render - LM Studio's filled container covers 69% of its
box against ~38% for the others, and Z.AI and OpenRouter are wide-but-short
marks spanning ~84% of the box height.

Choosing a target does not switch immediately. It runs the CLI's own preflight,
and if that refuses, the panel names what is missing and offers the fix - a
terminal for the hidden key prompt, a re-check once a server is up, or the
server form. If sessions are running it lists them by terminal and directory,
marks any mid-request, and asks whether to restart them, close them, or leave
them. The session action is applied strictly after the write, since restarting
first would only bring them back up on the provider just left.

The server form edits an LM Studio preset's base URL and whether it needs an API
key, reachable from the gear on any LM Studio preset row - per row, because two
presets can point at two different machines - and from the failure card. The
shipped local default is untouched unless it is changed.
 2026-08-30 21:07:03 +03:00
18 changed files with 3459 additions and 118 deletions
Expand all files Collapse all files
.gitignore
+3
View File
@@ -0,0 +1,3 @@
dist/
*.bak
*.bak.*
README.md
+453 -22
View File
@@ -2,7 +2,10 @@
Switch Claude Code system-wide between **Anthropic**, **OpenRouter**, **Z.AI**, and a Switch Claude Code system-wide between **Anthropic**, **OpenRouter**, **Z.AI**, and a
local **LM Studio** server — with named per-tier model presets. local **LM Studio** server — with named per-tier model presets.
Windows / PowerShell 5.1, no external dependencies.
Windows / PowerShell 5.1 with no external dependencies, and a POSIX port for
Linux and macOS that needs only python3 — including a theme-aware TUI and an
Omarchy bar widget. See [Linux / Omarchy](#linux--omarchy).
``` ```
claude-mode # interactive menu claude-mode # interactive menu
@@ -110,6 +113,13 @@ claude-mode preset rm <name>
claude-mode set-key [ref] store a key (hidden prompt, DPAPI) claude-mode set-key [ref] store a key (hidden prompt, DPAPI)
claude-mode models [filter] models available from the active provider claude-mode models [filter] models available from the active provider
claude-mode doctor verify auth, endpoint, model ids, stray env vars claude-mode doctor verify auth, endpoint, model ids, stray env vars
claude-mode preflight <mode> [preset] can this mode actually serve? (no switch)
claude-mode sessions running sessions, and which are mid-request
claude-mode sessions --stop close them (asks first)
claude-mode sessions --restart close and reopen each in its own directory
claude-mode sessions --dry-run show what either would do, and do nothing
claude-mode <mode> --force switch even when preflight says no
``` ```
Omitting the preset uses a **fixed** per-provider default, not "most recently Omitting the preset uses a **fixed** per-provider default, not "most recently
@@ -117,19 +127,24 @@ used" — so `claude-mode openrouter` always means `default`.
## Presets shipped ## Presets shipped
One per mode, so `claude-mode <mode>` is never ambiguous and there is no
menu to read before the thing you asked for happens. Build more with
`preset new` whenever one stops being enough.
| preset | provider | opus | sonnet | haiku | fable | | preset | provider | opus | sonnet | haiku | fable |
|---|---|---|---|---|---| |---|---|---|---|---|---|
| `default` | openrouter | `deepseek/deepseek-v4-flash` | `qwen/qwen3.7-flash` | `openrouter/free` | `z-ai/glm-5.2` | | `default` | openrouter | `z-ai/glm-5.3-flash` | `deepseek/deepseek-v4-flash-0731` | `openrouter/free` | `z-ai/glm-5.3` |
| `cheap` | openrouter | `deepseek/deepseek-v4-pro` | `deepseek/deepseek-v4-flash` | `qwen/qwen3.7-flash` | `openai/gpt-5.6-luna-pro` | | `zai` | zai | `glm-5.3` | `glm-5.3` | `glm-4.7` | `glm-5.3` |
| `claude-via-or` | openrouter | `anthropic/claude-opus-5` | `anthropic/claude-sonnet-5` | `anthropic/claude-haiku-4.5` | `anthropic/claude-fable-5` | | `lmstudio` | lmstudio | whatever setup finds on your server (all tiers) | | | |
| `zai` | zai | `glm-5.2` | `glm-5.2` | `glm-4.7` | `glm-5.2` |
| `lmstudio` | lmstudio | `kwaipilot_kat-coder-v2.5-dev` (all tiers) | | | |
| `lmstudio-qwen` | lmstudio | `qwen3.6-35b-a3b-uncensored-heretic-native-mtp-preserved` (all tiers) | | | |
Presets are plain JSON in `~/.claude-mode/presets/`. A preset declares its Presets are plain JSON in `~/.claude-mode/presets/`. A preset declares its
`provider`; `claude-mode lmstudio default` is rejected rather than silently `provider`; `claude-mode lmstudio default` is rejected rather than silently
pointing a local URL at remote model IDs. pointing a local URL at remote model IDs.
To route the real Anthropic models through OpenRouter, copy one and repoint the
tiers - `claude-mode preset new claude-via-or default`, then `preset set` each
tier to `anthropic/claude-opus-5` and friends.
## Context windows and early auto-compaction ## Context windows and early auto-compaction
**Symptom:** switch to a gateway and the session starts auto-compacting almost **Symptom:** switch to a gateway and the session starts auto-compacting almost
@@ -151,8 +166,8 @@ CLAUDE_CODE_AUTO_COMPACT_WINDOW = <contextTokens>
| preset | contextTokens | | preset | contextTokens |
|---|---| |---|---|
| `default`, `cheap`, `claude-via-or`, `zai` | 1,000,000 | | `default`, `zai` | 1,000,000 |
| `lmstudio`, `lmstudio-qwen` | 262,144 | | `lmstudio` | 262,144 |
`doctor` cross-checks the declared window against each tier's *actual* model `doctor` cross-checks the declared window against each tier's *actual* model
window and names any tier that falls short — `default` maps haiku to window and names any tier that falls short — `default` maps haiku to
@@ -233,17 +248,18 @@ Scanning your installed models' templates:
| **`qwen/qwen3.5-9b`**, **`prism-ml/bonsai-27b`** | **assert** | | **`qwen/qwen3.5-9b`**, **`prism-ml/bonsai-27b`** | **assert** |
`doctor` now reports this per model, and `models` flags affected entries with `doctor` now reports this per model, and `models` flags affected entries with
`TEMPLATE RISK`. **The fix is to use a model without the flag** — which is `TEMPLATE RISK`. **The fix is to use a model without the flag.** The Qwen3.6
exactly the `lmstudio-qwen` preset, verified end-to-end with a cold JIT load, line above is what the single `lmstudio` preset ships pointed at, verified
streaming, and tool calls. end-to-end with a cold JIT load, streaming, and tool calls.
Honest caveat: KAT-Coder's template *does* contain the assertion, but I could not Honest caveat: KAT-Coder's template *does* contain the assertion, but I could not
reproduce the failure against it here — cold JIT, streaming, tools, system reproduce the failure against it here — cold JIT, streaming, tools, system
blocks, and multi-turn `tool_result` all succeeded. Whether it trips seems to blocks, and multi-turn `tool_result` all succeeded. Whether it trips seems to
depend on which parser strategy the runtime picks. If it spams, switch: depend on which parser strategy the runtime picks. If the model setup lands on
does spam, run setup again and pick another:
```powershell ```bash
claude-mode lmstudio lmstudio-qwen claude-mode setup lmstudio
``` ```
Other notes: use a model with **>25k context** (`doctor` warns below that), and Other notes: use a model with **>25k context** (`doctor` warns below that), and
@@ -272,24 +288,148 @@ this tool writes is referenced by it:
## Setup ## Setup
One line, on any machine that can reach the repo (PowerShell 5.1 is fine):
```powershell ```powershell
cd c:\Users\smoido\Projects\cli\claude-code-switcher irm https://git.nebulm.com/smoido/claude-mode/raw/branch/master/install.ps1 | iex
```
Piped in like that there is no checkout, so the installer fetches the
repository archive to `%TEMP%` and installs from it. To pass switches
(`-Force`, `-SkipKeyPrompt`), download first and run the file:
```powershell
irm https://git.nebulm.com/smoido/claude-mode/raw/branch/master/install.ps1 -OutFile install.ps1
.\install.ps1 -Force
```
Or clone it — the normal path if you want to read the code and pull updates:
```powershell
git clone https://git.nebulm.com/smoido/claude-mode.git
cd claude-mode
.\install.ps1 .\install.ps1
``` ```
Installs to `~/.claude-mode/` (ACL: you only), drops `claude-mode.cmd` into All three install the same way: to `~/.claude-mode/` (ACL: you only), dropping
`~/.local/bin` (already on your User PATH, next to `claude.exe`), and adds a `claude-mode.cmd` into `~/.local/bin` (already on your User PATH, next to
marked block to `~/Documents/WindowsPowerShell/profile.ps1`. `claude.exe`), and adding a marked block to
`~/Documents/WindowsPowerShell/profile.ps1`.
`~/.claude/settings.json` is **not** touched by the installer — only by an actual `~/.claude/settings.json` is **not** touched by the installer — only by an actual
mode switch, which backs it up to `~/.claude-mode/backups/` first (last 20 kept). mode switch, which backs it up to `~/.claude-mode/backups/` first (last 20 kept).
Re-running any of them is a safe upgrade: scripts are overwritten, existing
presets are kept unless `-Force` is passed.
If the profile doesn't load: `Set-ExecutionPolicy -Scope CurrentUser RemoteSigned`. If the profile doesn't load: `Set-ExecutionPolicy -Scope CurrentUser RemoteSigned`.
## Restarting sessions ## Restarting sessions
Claude Code reads all of this **once at startup**. A switch does not affect a A switch **breaks** running sessions. This originally said it did not affect
running session — that's how the process loads config, not something scripting them; that was wrong, and the difference matters, because the two halves of the
can change. config behave differently.
The static half — base URL, model ids, the env block — genuinely is read once at
startup, and a running session keeps what it started with.
The credential is not. It comes from running `apiKeyHelper`, which Claude Code
re-invokes on a timer (`CLAUDE_CODE_API_KEY_HELPER_TTL_MS`, present in 2.1.251),
and the helper answers for whatever `state.json` says *at that moment*. So a
switch reaches into a live session through the one thing that was never cached:
| switching to | what the running session gets |
|---|---|
| `anthropic` | the helper returns nothing, by design — no credential at all |
| another provider | the new provider's key, sent to the old base URL, which rejects it |
| another preset of the *same* provider, same `keyRef` | same key, same endpoint — this one survives, on the model ids it started with |
Either of the first two starts failing calls whenever the TTL happens to expire:
mid-turn as easily as between turns.
### The part that is not just an inconvenience
A failed call is recoverable. A *successful* one may not be.
If a running session takes even one completion from the new provider before
anything notices — which happens when the mode it is switched to matches the
base URL it already had cached — that provider's message-id format lands in its
transcript. OpenRouter issues `gen-<epoch>-<rand>` where Anthropic issues
`msg_…`. Native Anthropic then refuses to resume the session at all:
```
API Error: 400 diagnostics.previous_message_id: must be the `id` from a
prior /v1/messages response (starts with `msg_`)
```
There is no supported way back from that. The transcript has to be rolled back
to the last message Anthropic issued:
```bash
claude-mode repair-session # transcripts for this project
claude-mode repair-session --all # every project, problems only
claude-mode repair-session <session-id> # show what it would cut
claude-mode repair-session <session-id> --apply
```
Scope, since it is not obvious: a bare listing covers only the project you are
standing in (walking up from the current directory to find it), while a **named
session id is looked up across every project** — you rarely remember which
project a session you cannot resume belonged to. `--all` drops the scoping
entirely.
`--all` reports only what is actually actionable, which matters more than it
sounds. Of 59 transcripts here it initially flagged 16; on inspection 5 had
simply never received a reply, and 10 had run start-to-finish on a gateway, so
every id in them is that provider's by design. Those resume perfectly well under
the provider they were born on, have nothing to truncate back to, and are only a
problem if you try to resume them as Anthropic. Neither is damage, so neither is
listed. Only a transcript that has a genuine `msg_` message *and* junk after it
is something this can or should touch.
### The cut turns are not thrown away
Truncating is the mechanical fix, but the turns being cut are the work itself —
losing the conversation that produced a morning's changes is most of the damage,
and a session that resumes with a hole in its memory is barely resumed at all.
So `--apply` does three things before it deletes anything:
1. **Backs up** the original as `<session>.jsonl.pre-repair-backup-<stamp>`.
2. **Writes the dropped turns out** as `<session>.recovered-<stamp>.md` — a
readable record of what was asked, what was answered, and what was run. Tool
*results* are left out; they are most of a transcript by volume and the least
useful part of a summary.
3. **Hands them back to the session** as a single appended note, so the agent
that resumes knows what it just did.
That note is a `user` entry marked `isMeta` — the same marker Claude Code uses
for its own local-command caveats, meaning "context, not something to answer".
Critically it carries **no `message.id`**, so it cannot re-create the very
condition being repaired. `--no-reinject` writes the Markdown but leaves the
session untouched.
It refuses to touch a transcript written to in the last 90 seconds, since that
one belongs to a session still alive.
### The bar notices for you
Nothing tells you a session is unresumable until you try to resume it, by which
point you have usually forgotten which one it was. So the widget scans every
project on a timer (and whenever the panel opens) and puts a dot on its icon
when there is something to fix. Clicking through lists the affected sessions and
offers to repair each one, after saying what it will drop and what it will keep.
The dot is a dot rather than a colour change, because recolouring the mark would
misreport the active mode — which is the widget's actual job.
That scan is only affordable because it reads the *tail* of each transcript
first: if the last message is Anthropic's, the transcript is healthy and the
rest of the file is never opened. Since that is the overwhelmingly common case,
the whole sweep costs ~60ms for 59 transcripts, against ~5s for the obvious
version that reads every byte of every one.
This is why a switch now asks before it writes rather than reporting afterwards.
So restart affected sessions — and `claude-mode sessions --restart` will do it
for you:
- **CLI** — exit and relaunch `claude` - **CLI** — exit and relaunch `claude`
- **VS Code** — `Ctrl+Shift+P` → *Developer: Reload Window* - **VS Code** — `Ctrl+Shift+P` → *Developer: Reload Window*
@@ -316,6 +456,297 @@ can change.
`extraEnv` key (Z.AI's timeouts, LM Studio's attribution header) is removed when `extraEnv` key (Z.AI's timeouts, LM Studio's attribution header) is removed when
you switch away — even though no other preset knows that key exists. you switch away — even though no other preset knows that key exists.
## Linux / Omarchy
The `linux/` tree is a POSIX port of the same design: one write to
`~/.claude/settings.json`, secrets kept out of it, presets shared verbatim with
the Windows build.
```bash
cd claude-code-switcher
bash linux/install.sh # --skip-key-prompt to install without storing a key
```
Installs to `~/.claude-mode/`, symlinks `~/.local/bin/claude-mode`, and adds the
`claude` wrapper to `~/.bashrc` / `~/.zshrc` between markers.
Secrets have no DPAPI equivalent here, so the vault picks the best backend
available and says which one it chose: macOS Keychain, libsecret, `pass`, or a
0600 file that is honestly labelled as unencrypted.
### The menu follows your desktop theme
The sixteen ANSI colour slots carry no guarantee about relative brightness, and
monochrome themes exploit that. Under Omarchy's Solitude, slot 36 — headings —
resolves to `#707070` and slot 31 — `FAIL` — to `#565d60`. Against `#cacccc`
body text on a `#101315` ground that is 3.8:1 and 2.8:1 where the body text is
11.6:1, so headings render as fine print and an error becomes the quietest
thing on screen. Exactly backwards.
So when Omarchy is present, the palette is derived from the theme it publishes
at `~/.local/state/omarchy/current/theme/colors.toml` instead. Every role is
measured against the background it will actually be drawn on and lifted toward
the foreground when it falls short, which keeps hue where the theme has any and
falls back to weight where it does not:
| role | before (Solitude) | after |
|---|---|---|
| heading | 3.8:1 | 9.4:1 |
| `FAIL` | 2.8:1 | 5.2:1 |
| help text | 2.2:1 | 2.2:1 (recessive on purpose, floored) |
Light themes are handled by the same arithmetic — `mode` in `colors.toml` is
authoritative, so `catppuccin-latte` and `flexoki-light` keep their accents
rather than washing out.
Overrides: `CLAUDE_MODE_THEME=/path/to/colors.toml` points it elsewhere,
`NO_COLOR` turns it off. Without Omarchy, or on a terminal that cannot do
truecolor, it falls back to the ANSI slots with the two roles the slots get
wrong corrected — bright red for `FAIL`, bold on headings.
### `claude-mode health`
Rewrites `~/.claude-mode/health.json`, the machine-readable mirror of the
active configuration: mode, preset, model map, context window, key backend, and
the switchable preset catalogue. No key material. A switch and a `status` both
refresh it; the command exists for anything that wants to force it.
## First-run setup
A shipped preset is a starting point, not a working configuration. OpenRouter
and Z.AI have no key stored. LM Studio's model ids are whatever happened to be
installed on the machine this was packaged on, which is almost certainly not
yours.
So the shipped presets carry `configured: false`, and preflight treats that as a
blocker with its own remedy:
```bash
claude-mode setup openrouter # key, then models from OpenRouter's catalogue
claude-mode setup zai # key, then per-tier GLM models
claude-mode setup lmstudio # server URL, auth, then models from that server
claude-mode setup anthropic # nothing to do; it uses your existing login
```
Setup asks only what it cannot work out, shows the current model map before
offering to change it, and picks from the provider's own catalogue rather than
asking anyone to type a model id from memory. LM Studio maps one model across
every tier, since a local server has one loaded at a time and per-tier mapping
would just pay the load cost on every tier change; the remote gateways ask per
tier, which is the point of them.
In a terminal, a switch that trips this offers to run setup there and then
rather than printing a command to type next. It sets `configured: true` on the
way out.
**Absent means configured.** Presets that predate this, and any built by hand
with `preset new`, are nobody's business but yours and will not start demanding
a wizard.
## Refusing a switch that would not work
A switch writes settings.json and is picked up by the *next* `claude` launch, so
switching into a mode that cannot serve requests does not fail loudly — it
succeeds, and every session started afterwards is broken in a way that points at
Claude Code rather than at here.
LM Studio is the sharp case. Its token is an inline placeholder, so nothing
about the switch needs the server to exist; point at a server that is not
running and you get a config that looks perfectly healthy and answers nothing.
So the preconditions are checked before the write, not after:
| mode | checked | when it fails |
|---|---|---|
| any gateway | the preset has been through setup at least once | `claude-mode setup <mode>` |
| `anthropic` | nothing to check | — |
| `openrouter`, `zai` | a key exists for the preset's `keyRef`, and the helper is executable | `claude-mode set-key <ref>` |
| `lmstudio` (anywhere) | the server answers, and accepts the credential the preset would send | see below |
| any preset on a local base URL | same probe | start the server |
| all | the preset exists and declares the provider being switched to | — |
The probe distinguishes four outcomes, because their remedies are opposites:
| result | means | remedy offered |
|---|---|---|
| `ok` | answered `/api/v0/models` or `/v1/models` | — |
| `auth` | the server is up and refused the credential | store or fix the key, or turn auth on for the preset |
| `notfound` | something is listening, but the API is not at that path | fix the base URL |
| `refused` | nothing answered at all — down, asleep, DNS, TLS, timeout | start the server, or fix the address |
`claude-mode preflight <mode> [preset]` runs exactly these and prints the
verdict as JSON without switching; it is what the bar widget calls before it
offers to do anything. `--force` overrides the lot.
Remote *gateways* are not probed — OpenRouter or Z.AI being briefly unreachable
is the network's problem and not worth blocking a config change over, where a
missing key never fixes itself. LM Studio is probed wherever it lives, because
an instance on a sleeping LAN box is exactly as absent as a loopback port with
nothing behind it, and produces the identical silent breakage. Off-machine
addresses get a longer timeout, not a pass.
### LM Studio somewhere other than this machine
It ships on loopback, but that is a default, not a constraint. A preset is just
a base URL and an auth block, so all of these are the same two fields:
```bash
claude-mode preset url lmstudio http://192.168.1.40:1234 # another box on the LAN
claude-mode preset url lmstudio https://lms.example.net # through a tunnel or proxy
claude-mode preset auth lmstudio key lmstudio # that server wants an API key
claude-mode set-key lmstudio # store it (vault, not settings.json)
claude-mode preset auth lmstudio none # back to the open-server default
```
`auth none` writes LM Studio's inline placeholder token, which is not a secret
and is what an unauthenticated server expects. `auth key` moves it to the vault
like every other credential — a real key on a public address is a real key.
Nothing here changes the shipped presets unless you ask it to; the local default
stays exactly as it was.
## Sessions still on the old provider
A switch does not leave running sessions on the old provider — it breaks them.
See [Restarting sessions](#restarting-sessions) for the mechanism: their
endpoint is fixed at startup but their credential is re-fetched on a timer, so
it switches under them and the endpoint they are still pointed at refuses it.
Which is why this is a decision rather than a notification.
A switch stops and asks while any are running, offering to restart them (the
only answer that ends with everything on the mode the bar now claims), close
them, proceed anyway, or abort — and abort is the default. Non-interactively it
refuses outright unless given `--yes`.
```
claude-mode sessions
running sessions (2)
562250 pts/4 /home/smoido/Work working (this session - never touched)
631644 pts/1 /home/smoido/Projects/api
```
Sessions are found through `/proc/<pid>/exe`, which on Linux resolves to the
real `claude` binary — a process-name match would sweep up every shell that
merely mentions claude on its command line, including the one this is running
from. Two things are then filtered out:
- **The calling session.** Killing the session that asked for the kill is not a
thing anyone means, so it is listed and never signalled.
- **Forks of a session.** A busy session spawns children off its own binary,
and they inherit the same `exe`. Without excluding anything whose parent is
itself claude, the count climbs and falls with how hard the machine is
thinking — it read 2, 5, 11 and 40 on the same two sessions before this. A
real session's parent is a terminal.
`working` is a sampled-CPU heuristic — two reads of `utime + stime` 300ms apart
— so it is a good guess about which session is mid-turn, not a promise.
`--stop` sends SIGTERM (never SIGKILL; Claude Code writes out its transcript on
the way down). `--restart` stops each session and reopens it — re-running the
parent terminal's own command line where there is one, so the same terminal,
flags and directory come back, and falling back to a fresh terminal in the
session's directory otherwise. Interactive runs confirm first; `--yes` is for
callers that have already asked, and `--dry-run` prints the plan and touches
nothing.
## The Omarchy bar widget
An icon in the Omarchy top bar showing which provider the next `claude` launch
will use, and a panel that switches it without a terminal.
```bash
bash omarchy/install.sh
```
Copies the plugin to `~/.config/omarchy/plugins/smoido.claude-mode/` and adds
its id to the bar layout in `~/.config/omarchy/shell.json` (backed up first).
Both hot-reload, so nothing needs restarting — except after a change to
`Modes.js`, which the QML engine caches for the life of the process as a
`.pragma library` (`omarchy restart shell`).
The icon is the mode, and it is the provider's own logo: the Claude burst, the
OpenRouter arrow, the Z.AI Z, the LM Studio mark. They are drawn as vector paths
with `QtQuick.Shapes` rather than set as font glyphs — three of the four have no
Nerd Font pictograph at all, and being paths means they take the bar's
foreground colour and follow the theme like everything else. Anything other than
Anthropic takes the theme accent, so the bar stays quiet exactly when nothing
unusual is configured.
Marks come from [simple-icons](https://simpleicons.org) (Claude, OpenRouter, LM
Studio) and [lobe-icons](https://github.com/lobehub/lobe-icons) (Z.AI);
trademarks belong to their owners. Each carries an optical scale factor, because
equal nominal size is not equal apparent size — LM Studio's filled container
covers 69% of its box against ~38% for the other three, and Z.AI and OpenRouter
are wide-but-short marks whose ink spans only ~84% of the box height. The
factors in `Modes.js` are the geometric mean of both corrections.
Three details keep them from looking ragged in a 13px slot, which is what every
stock glyph in this bar measures:
- **No layer.** A layer rasterises the Shape at its own size and then scales the
*texture*, so a 24px buffer minified to 13 resamples ~2 pixels into 1. Without
one the scale is a transform on the geometry and rasterisation happens once,
at final resolution.
- **`Shape.CurveRenderer`** (Qt 6.6+) rasterises curves analytically instead of
tessellating them into antialiased triangles. Measured against a cairo render
of the same mark at the same size, the two now come out identical.
- **An odd `iconSize`.** These marks are radially symmetric, so their vertical
and horizontal arms sit on the centre line — which is a pixel *centre* at an
odd size and the seam between two pixels at an even one, where each arm splits
its coverage and greys out. Even values are rounded up.
- **Click** the icon for the panel: current mode, every mode with the active one
ticked, and the model map behind a gateway. Picking a gateway unfolds its
presets rather than switching blind; picking a preset starts the switch.
Choosing a target does not switch immediately — it runs the same two checks the
CLI does, and both can stop it:
1. **Preflight.** If the mode has not been set up, has no key stored, or its
server is not answering, the panel says which and offers the fix: *Set up
<mode>…* opens a terminal and runs the whole first-run flow there (a bar popup
can host neither a hidden key prompt nor a filter-select model list), *Store
the key…* opens a terminal for just the prompt, *Check again* re-runs the
preflight, and *Server settings…* opens the form below. The gear on any LM Studio preset row opens the same form without
waiting for a failure — per row, because two LM Studio presets can point at
two different machines.
The form holds the base URL, a *Use local default* reset, and a switch for
whether that server needs an API key. Saving rewrites the preset and drops
straight back into the switch that was blocked.
2. **Running sessions.** If any are running, the panel lists them by terminal
and directory, marks any that are mid-request, and asks: *Switch and restart*,
*Switch and close*, *Switch only*, or *Cancel*. The session action is applied
strictly **after** the write — restarting first would only bring them back up
on the provider you just left.
- **Right-click** switches straight back to Anthropic.
- **Middle-click** re-reads state.
- **Hover** for mode, preset, and the opus/sonnet mapping.
State comes from watching `health.json`, not from polling the CLI, so the
widget costs nothing while idle and a switch made in a terminal shows up in the
bar on its own. A failed switch — a preset whose key was never stored is the
common one — surfaces the CLI's own error in the panel rather than looking
like a click that did nothing.
Per-instance settings in the `shell.json` layout entry:
| key | default | meaning |
|---|---|---|
| `showLabel` | `false` | show the preset name beside the icon as well |
| `iconSize` | `13` | mark size in px; rounded up to odd (see below) |
| `root` | `~/.claude-mode` | where claude-mode is installed |
Placement is `right`, before `omarchy.agents`; override with `CM_BAR_SECTION`
and `CM_BAR_BEFORE` when installing. Moving it later is a normal
`omarchy bar move smoido.claude-mode --section <section>`.
Uninstall:
```bash
rm -rf ~/.config/omarchy/plugins/smoido.claude-mode
# then remove the {"id": "smoido.claude-mode"} entry from ~/.config/omarchy/shell.json
```
## Uninstall ## Uninstall
```powershell ```powershell
claude-mode.ps1
+2 -2
View File
@@ -1251,7 +1251,7 @@ function Invoke-Models {
if ($mode -eq 'zai') { if ($mode -eq 'zai') {
Write-Head 'Z.AI GLM models (from Z.AI docs - no public catalogue endpoint)' Write-Head 'Z.AI GLM models (from Z.AI docs - no public catalogue endpoint)'
@('glm-5.2 - flagship coding model (opus/sonnet tier)', @('glm-5.3 - flagship coding model (opus/sonnet tier)',
'glm-4.7 - fast/cheap tier (haiku tier)') | ForEach-Object { Write-Host " $_" } 'glm-4.7 - fast/cheap tier (haiku tier)') | ForEach-Object { Write-Host " $_" }
Write-Host ' Full list: https://docs.z.ai/devpack/tool/claude' Write-Host ' Full list: https://docs.z.ai/devpack/tool/claude'
return return
@@ -1856,7 +1856,7 @@ function Get-ModelChoices {
elseif ($provider -eq 'zai') { elseif ($provider -eq 'zai') {
# Z.AI publishes no public catalogue endpoint; these are the documented # Z.AI publishes no public catalogue endpoint; these are the documented
# coding-plan models. Anything else can still be typed manually. # coding-plan models. Anything else can still be typed manually.
$out += [pscustomobject]@{ Id = 'glm-5.2'; Key = 'glm-5.2'; Label = 'glm-5.2'; Detail = @('flagship coding model - opus/sonnet tier') } $out += [pscustomobject]@{ Id = 'glm-5.3'; Key = 'glm-5.3'; Label = 'glm-5.3'; Detail = @('flagship coding model - opus/sonnet tier') }
$out += [pscustomobject]@{ Id = 'glm-4.7'; Key = 'glm-4.7'; Label = 'glm-4.7'; Detail = @('fast/cheap tier - haiku') } $out += [pscustomobject]@{ Id = 'glm-4.7'; Key = 'glm-4.7'; Label = 'glm-4.7'; Detail = @('fast/cheap tier - haiku') }
} }
docs/incident-mode-switch-corrupts-live-sessions.md
+142
View File
@@ -0,0 +1,142 @@
# Incident: switching modes mid-session corrupts and can permanently break that session
**Date:** 2026-08-30/31
**Affected session:** `1fffec9e-e2e3-4255-828c-a9ccfd8f631d` (project `/home/smoido/Work`)
**Severity:** High — silent, mid-task corruption with no warning, and one failure path is not recoverable by normal means (requires manual transcript surgery).
## Summary
The user had a live, running Claude Code session open in `~/Work`. While that
session was mid-task, `claude-mode` was used to switch the active provider
away from `anthropic` and later back to `anthropic`. The running session was
not restarted in between. This corrupted the session's credentials, and — because
the session had picked up a different provider's message-ID format in the
meantime — it became permanently unable to resume under native Anthropic auth,
throwing a hard API error on every resume attempt.
This must not be possible to trigger silently. Whatever ships next needs to
either prevent it, warn loudly before it happens, or make recovery automatic.
## Root cause / mechanism
`claude-mode` writes provider config into `~/.claude/settings.json`
(`env` block: base URL, model IDs, etc., plus `apiKeyHelper` pointing at
`claude-key-helper.sh`). A **running** `claude` process only reads part of
this at startup:
- **Static, cached at startup:** `ANTHROPIC_BASE_URL`, the model ID env vars,
the rest of the `env` block. A running session keeps whatever it started
with here — switching modes does *not* change these for an
already-running process.
- **Not cached — re-fetched on a timer:** the credential. Claude Code
re-invokes `apiKeyHelper` periodically (`CLAUDE_CODE_API_KEY_HELPER_TTL_MS`
is present in the 2.1.251 binary), and `claude-key-helper.sh` answers based
on whatever `~/.claude-mode/state.json` says *at the moment it's called* —
not at session startup.
So a mode switch reaches into a live session through the one part that was
never cached:
| switching the global mode to... | what the *running* session gets on next credential refresh |
|---|---|
| `anthropic` | the helper returns nothing (native anthropic auth is expected to need no helper) → session has no credential at all |
| any other provider/preset | the new key, but the session is still pointed at the **old** base URL (cached at startup) → that endpoint rejects the new key |
Either way, the session starts failing API calls from the moment the TTL
next expires — mid-turn as easily as between turns. The one case that does
*not* break: switching between two presets of the *same* provider that share
a `keyRef` (same key, same endpoint survives).
`claude-mode status` and `claude-mode health` only ever report the *global*
config's current state. Neither one has any way to know a specific already-running
session exists, let alone that it's about to be (or has been) knocked over.
## What happened to this specific session (concrete failure chain)
1. Session `1fffec9e` was live in `~/Work`, working on `claude-code-switcher`
itself.
2. User ran a switch to the `openrouter` "default" preset (via `claude-mode`)
to test something, while that session kept running.
3. The running session's next few turns actually succeeded against
openrouter (its cached base URL now matched, since the switch happened to
land on the provider its env pointed at) — those turns got real
completions back, but with **openrouter's message-ID format**
(`gen-<timestamp>-<id>`), not Anthropic's (`msg_<id>`).
4. A later call failed outright and Claude Code recorded a synthetic
client-side placeholder turn (`"model":"<synthetic>"`,
`"isApiErrorMessage":true`, `"error":"unknown"`) as the last message in
the transcript — this is the visible "session disconnected mid-task"
symptom.
5. User switched the global mode back to `anthropic`.
6. Resuming session `1fffec9e` (`claude -r 1fffec9e-... -p "..."`) now fails
unconditionally with:
```
API Error: 400 diagnostics.previous_message_id: must be the `id` from a
prior /v1/messages response (starts with `msg_`)
```
because native Anthropic's API requires `previous_message_id` to be a
real Anthropic-issued ID, and the last message(s) in this session's
transcript are not (`gen-...` or the synthetic placeholder ID).
7. **There is no supported way to resume past this.** The only fix found was
manual surgery on the session's `.jsonl` transcript file: locate the last
message that still has a genuine `msg_...` id (in this case, several
turns earlier, at a clean `end_turn` boundary well before the switch was
even tested), back up the original file, and truncate the transcript to
that point. That rolls the session back to its last-known-good state and
makes it resumable again — at the cost of permanently losing every turn
after that point (the actual code changes from that later work were
separately safe in git, but the chat narrative was not recoverable).
## Impact
- Silent corruption: nothing warns the user before or during the switch that
a live session exists and is about to break.
- Depending on timing, the break can be "just" an auth failure (annoying,
session still resumable once you're back on the mode it started with) **or**
a hard, unrecoverable-by-normal-means failure (if the session round-tripped
through a different provider's ID format before failing) that requires
hand-editing a JSONL transcript to fix.
- This applies to *every* running `claude` process on the machine at switch
time, not just the one in the foreground shell — CLI, VS Code, desktop, any
of them.
## Requirements for the fix
1. **Detect running sessions before switching.** `claude-mode` should scan
for live `claude` processes (and ideally which project/cwd each belongs
to) before performing a switch.
2. **Warn or block, don't silently proceed.** At minimum, print a clear
warning naming the affected session(s)/PIDs/cwds and what will happen to
them (credential will be pulled out from under them on the next TTL
refresh). Consider requiring `--force` (or an explicit confirmation) to
proceed while sessions are running, and defaulting to "abort" otherwise.
3. **Prefer a safe path when sessions are detected:** e.g. offer to let the
user gracefully end/save those sessions first, or clearly instruct them
to restart affected sessions immediately after the switch completes.
4. **Document the real mechanism** (this file's "Root cause" section) in the
tool's own help/README, replacing any prior claim that a switch "does not
affect a running session" or that sessions "keep talking to the old
provider until restarted" — both are wrong; the credential moves under
them regardless.
5. **Make the unrecoverable failure mode recoverable.** Add a `claude-mode
repair` (or similar) command that:
- finds a given session's transcript,
- locates the last message with a valid `msg_...` id,
- backs up the original file,
- truncates to that point,
so this doesn't require manual `jq`/`head`/`grep` surgery next time.
This is the exact procedure used to fix session `1fffec9e` above.
## Repro steps (for verification once fixed)
1. Start a `claude` session in some project directory, mid-task.
2. In a separate shell, run `claude-mode <some-other-provider>` (something
with a different base URL/key format from the session's current mode).
3. Let the running session's next API-key TTL refresh happen (or just issue
another prompt in it) — observe the call fail.
4. Switch back: `claude-mode anthropic`.
5. Try `claude -r <that-session-id> -p "hi"` — currently fails with the
`previous_message_id` 400 error if the session ever got a non-`msg_`
completion in between. The fix should make step 2 impossible (or clearly
confirmed) rather than needing step 5's failure to be caught after the
fact.
install.ps1
+27 -1
View File
@@ -18,7 +18,33 @@ param(
Set-StrictMode -Version 1.0 Set-StrictMode -Version 1.0
$ErrorActionPreference = 'Stop' $ErrorActionPreference = 'Stop'
$src = Split-Path -Parent $MyInvocation.MyCommand.Path $repoUrl = 'https://git.nebulm.com/smoido/claude-mode'
# Running from a checkout, the source is the files next to this script. Piped
# in (irm ... | iex) there is no script file and no checkout, so fetch the
# repository archive and install from that instead.
$src = $null
if ($PSScriptRoot -and (Test-Path -LiteralPath (Join-Path $PSScriptRoot 'claude-mode.ps1'))) {
$src = $PSScriptRoot
}
else {
[Net.ServicePointManager]::SecurityProtocol =
[Net.ServicePointManager]::SecurityProtocol -bor [Net.SecurityProtocolType]::Tls12
$tmp = Join-Path $env:TEMP ('claude-mode-src-' + [IO.Path]::GetRandomFileName().Replace('.', ''))
New-Item -ItemType Directory -Path $tmp -Force | Out-Null
$zip = Join-Path $tmp 'repo.zip'
Write-Host "no local checkout found - fetching source from $repoUrl" -ForegroundColor Cyan
Invoke-WebRequest -Uri "$repoUrl/archive/master.zip" -OutFile $zip -UseBasicParsing
Expand-Archive -LiteralPath $zip -DestinationPath $tmp
# Forgejo nests everything under a repo-named folder; find it wherever it is.
$extracted = Get-ChildItem -LiteralPath $tmp -Directory |
Where-Object { Test-Path -LiteralPath (Join-Path $_.FullName 'claude-mode.ps1') } |
Select-Object -First 1
if ($extracted) { $src = $extracted.FullName }
elseif (Test-Path -LiteralPath (Join-Path $tmp 'claude-mode.ps1')) { $src = $tmp }
else { throw 'archive did not contain claude-mode.ps1 - repo layout changed?' }
}
$root = Join-Path $env:USERPROFILE '.claude-mode' $root = Join-Path $env:USERPROFILE '.claude-mode'
Write-Host "installing claude-mode -> $root" -ForegroundColor Cyan Write-Host "installing claude-mode -> $root" -ForegroundColor Cyan
linux/claude-mode
+564 -34
View File
@@ -274,10 +274,15 @@ claude-mode - switch Claude Code between Anthropic, OpenRouter, Z.AI, LM Studio
claude-mode doctor verify auth, endpoint, model ids, env claude-mode doctor verify auth, endpoint, model ids, env
claude-mode repair strip [1m] tags from cached model ids claude-mode repair strip [1m] tags from cached model ids
claude-mode health refresh health.json (machine-readable state) claude-mode health refresh health.json (machine-readable state)
claude-mode setup <mode> [--terminal] first-run setup: key, server, models
claude-mode preflight <mode> [preset] check a mode can actually serve, without switching claude-mode preflight <mode> [preset] check a mode can actually serve, without switching
claude-mode sessions [--stop|--restart] claude-mode sessions [--stop|--restart]
running sessions; close or reopen them running sessions; close or reopen them
claude-mode repair-session [id] [--apply]
make a session resumable again after a bad switch,
keeping the cut turns as markdown + a context note
claude-mode <mode> --force switch even if preflight says no claude-mode <mode> --force switch even if preflight says no
claude-mode <mode> --yes switch without asking about running sessions
EOF EOF
} }
@@ -453,6 +458,13 @@ cm_preflight() {
return 1 return 1
fi fi
if ! preset_configured "$pf"; then
cm_pf_set 'needs-setup' "$(mode_label "$mode" | cut -d- -f1 | sed 's/ *$//') has not been set up yet" \
"The shipped preset is a starting point: it has no key stored, and its model ids are whatever was on the machine this was packaged on. Setup asks for what it needs and picks models from the provider's own catalogue." \
"claude-mode setup $mode" 'setup'
return 1
fi
base="$(jget "$pf" baseUrl)"; CM_PF_BASEURL="$base" base="$(jget "$pf" baseUrl)"; CM_PF_BASEURL="$base"
auth_mode="$(jget "$pf" auth.mode)"; [ -z "$auth_mode" ] && auth_mode=vault auth_mode="$(jget "$pf" auth.mode)"; [ -z "$auth_mode" ] && auth_mode=vault
@@ -539,14 +551,31 @@ cmd_preflight() {
# --------------------------------------------------------------------------- # ---------------------------------------------------------------------------
# Running sessions # Running sessions
# #
# Claude Code reads settings.json once, at startup. A switch therefore does # A switch breaks running sessions. Not "leaves them on the old provider" -
# nothing to a session already running - it keeps talking to the old provider on # breaks them, and it is worth being exact about why, because the two halves of
# the old key until it is restarted, which is the confusing part: the bar says # the config behave differently.
# one thing and the session in front of you is doing another.
# #
# Worse for a session mid-request. The key it is using can be pulled out from # The static half - base URL, model ids, the env block - really is read once at
# under it (anthropic mode deletes the helper outright), so an in-flight turn # startup, and a running session keeps the values it started with.
# can fail on the next tool call rather than at a clean boundary. #
# The credential is not. It is fetched by running apiKeyHelper, which Claude
# Code re-invokes on a timer (CLAUDE_CODE_API_KEY_HELPER_TTL_MS, present in
# 2.1.251), and the helper answers for whatever state.json says *now*. So a
# switch reaches into a live session through the one thing that was never
# cached:
#
# -> anthropic the helper returns nothing at all, by design, and the
# session's next refresh comes back with no credential
# -> another provider
# the helper hands over the new provider's key while the
# session is still pointed at the old base URL, which rejects
# it
#
# Either way the session starts failing its calls, at whatever moment the TTL
# happens to expire - mid-turn as easily as between turns. The one case that
# does survive is a switch between two presets of the same provider sharing a
# keyRef: same key, same endpoint, and the session simply carries on with the
# model ids it started with.
# #
# Sessions are found through /proc/<pid>/exe rather than by matching process # Sessions are found through /proc/<pid>/exe rather than by matching process
# names. `claude` is a real ELF binary here, so the symlink resolves to it # names. `claude` is a real ELF binary here, so the symlink resolves to it
@@ -690,7 +719,7 @@ $rows
EOF_ROWS EOF_ROWS
if [ "$action" = "list" ]; then if [ "$action" = "list" ]; then
printf '\n %sthese keep the provider they started with until restarted%s\n' "$C_DIM" "$C_RESET" printf '\n %stheir credential is re-fetched on a timer, so a switch breaks them%s\n' "$C_DIM" "$C_RESET"
printf ' %sclaude-mode sessions --stop close them%s\n' "$C_DIM" "$C_RESET" printf ' %sclaude-mode sessions --stop close them%s\n' "$C_DIM" "$C_RESET"
printf ' %sclaude-mode sessions --restart close and reopen each in its own directory%s\n' "$C_DIM" "$C_RESET" printf ' %sclaude-mode sessions --restart close and reopen each in its own directory%s\n' "$C_DIM" "$C_RESET"
printf ' %s--dry-run shows what either would do, and does nothing%s\n' "$C_DIM" "$C_RESET" printf ' %s--dry-run shows what either would do, and does nothing%s\n' "$C_DIM" "$C_RESET"
@@ -793,24 +822,6 @@ EOF_ROWS
return 0 return 0
} }
# Named after the switch, not before it: the switch has already happened, and
# these are the sessions it did not reach.
cm_report_live_sessions() {
local rows n busy
rows="$(cm_session_rows)"
n="$(printf '%s' "$rows" | grep -c . || true)"
[ "${n:-0}" -gt 0 ] || return 0
busy="$(printf '%s' "$rows" | cut -f4 | grep -c '^yes$' || true)"
printf '\n'
warn "$n Claude Code session(s) still running on the previous provider"
if [ "${busy:-0}" -gt 0 ]; then
warn "$busy of them is mid-request - it may fail on its next call rather than at a clean stop"
fi
printf ' %sclaude-mode sessions what is running%s\n' "$C_DIM" "$C_RESET"
printf ' %sclaude-mode sessions --restart close and reopen them on the new provider%s\n' "$C_DIM" "$C_RESET"
}
cm_terminal_cmd() { cm_terminal_cmd() {
local t local t
for t in "${TERMINAL:-}" foot alacritty ghostty kitty; do for t in "${TERMINAL:-}" foot alacritty ghostty kitty; do
@@ -849,13 +860,30 @@ set_mode() {
if [ "$CM_FORCE" -eq 0 ] && ! cm_preflight "$mode" "$preset_name"; then if [ "$CM_FORCE" -eq 0 ] && ! cm_preflight "$mode" "$preset_name"; then
err "$CM_PF_TITLE" err "$CM_PF_TITLE"
[ -n "$CM_PF_DETAIL" ] && say "$CM_PF_DETAIL" [ -n "$CM_PF_DETAIL" ] && say "$CM_PF_DETAIL"
# Standing in a terminal with the fix one keystroke away, printing
# the command to type next is a poor substitute for running it.
if ui_interactive && [ "$CM_PF_KIND" = "setup" ]; then
printf '\n'
if ask_yes "set up $mode now?"; then
cmd_setup "$mode" "$preset_name" || return 1
cm_preflight "$mode" "$preset_name" || {
err "$CM_PF_TITLE"; return 1
}
else
return 1
fi
else
[ -n "$CM_PF_REMEDY" ] && printf ' %sfix:%s %s\n' "$C_DIM" "$C_RESET" "$CM_PF_REMEDY" [ -n "$CM_PF_REMEDY" ] && printf ' %sfix:%s %s\n' "$C_DIM" "$C_RESET" "$CM_PF_REMEDY"
printf ' %s--force switches anyway%s\n' "$C_DIM" "$C_RESET" printf ' %s--force switches anyway%s\n' "$C_DIM" "$C_RESET"
return 1 return 1
fi fi
fi
[ -f "$preset_file" ] || { err "preset '$preset_name' not found"; return 1; } [ -f "$preset_file" ] || { err "preset '$preset_name' not found"; return 1; }
fi fi
cm_confirm_sessions "$mode" || return 1
backup="$(backup_settings)" backup="$(backup_settings)"
if ! "$PY" "$JSON" apply "$CM_SETTINGS" "$CM_STATE" "$mode" "$preset_file" "$CM_HELPER" >/dev/null; then if ! "$PY" "$JSON" apply "$CM_SETTINGS" "$CM_STATE" "$mode" "$preset_file" "$CM_HELPER" >/dev/null; then
@@ -903,7 +931,7 @@ set_mode() {
check_stale_models "$mode" check_stale_models "$mode"
write_health "$mode" "$preset_name" write_health "$mode" "$preset_name"
cm_report_live_sessions cm_apply_session_action
printf '\n %srestart claude (and reload the VS Code window) to pick this up%s\n' "$C_DIM" "$C_RESET" printf '\n %srestart claude (and reload the VS Code window) to pick this up%s\n' "$C_DIM" "$C_RESET"
} }
@@ -1088,9 +1116,26 @@ or_catalogue() {
curl -fsS --max-time 30 https://openrouter.ai/api/v1/models 2>/dev/null | "$PY" "$JSON" or-models 2>/dev/null curl -fsS --max-time 30 https://openrouter.ai/api/v1/models 2>/dev/null | "$PY" "$JSON" or-models 2>/dev/null
} }
# The credential a preset would send. Empty when there is nothing to send.
cm_preset_token() {
local pf="$1" am ref
am="$(jget "$pf" auth.mode)"; [ -z "$am" ] && am=vault
if [ "$am" = "vault" ]; then
ref="$(jget "$pf" auth.keyRef)"; [ -z "$ref" ] && ref=openrouter
cm_vault_get "$ref" 2>/dev/null || true
else
jget "$pf" auth.token
fi
}
# The token is not optional decoration. An LM Studio server with authentication
# switched on answers /api/v0/models with 401 like anything else, so without it
# the catalogue comes back empty and every caller silently believes the server
# has no models installed - on exactly the setups that need the list most.
lms_catalogue() { lms_catalogue() {
local base="${1%/}" local base="${1%/}" token="${2:-}"
curl -fsS --max-time 10 "$base/api/v0/models" 2>/dev/null | "$PY" "$JSON" lms-models 2>/dev/null curl -fsS --max-time 10 ${token:+-H "Authorization: Bearer $token"} \
"$base/api/v0/models" 2>/dev/null | "$PY" "$JSON" lms-models 2>/dev/null
} }
# --------------------------------------------------------------------------- # ---------------------------------------------------------------------------
@@ -1162,14 +1207,14 @@ cmd_models() {
case "$mode" in case "$mode" in
lmstudio) lmstudio)
head_ "models installed in LM Studio at $(jget "$pf" baseUrl)" head_ "models installed in LM Studio at $(jget "$pf" baseUrl)"
lms_catalogue "$(jget "$pf" baseUrl)" | while IFS=$'\t' read -r id st ctx; do lms_catalogue "$(jget "$pf" baseUrl)" "$(cm_preset_token "$pf")" | while IFS=$'\t' read -r id st ctx; do
[ -z "$filter" ] || case "$id" in *"$filter"*) ;; *) continue ;; esac [ -z "$filter" ] || case "$id" in *"$filter"*) ;; *) continue ;; esac
printf ' %-58s %-11s %s\n' "$id" "$st" "$ctx" printf ' %-58s %-11s %s\n' "$id" "$st" "$ctx"
done done
;; ;;
zai) zai)
head_ 'Z.AI GLM models (from Z.AI docs - no public catalogue endpoint)' head_ 'Z.AI GLM models (from Z.AI docs - no public catalogue endpoint)'
say 'glm-5.2 - flagship coding model (opus/sonnet tier)' say 'glm-5.3 - flagship coding model (opus/sonnet tier)'
say 'glm-4.7 - fast/cheap tier (haiku tier)' say 'glm-4.7 - fast/cheap tier (haiku tier)'
;; ;;
*) *)
@@ -1276,7 +1321,7 @@ else: print(' ok spend %.2f of %.2f limit (%s), %.2f remaining' % (use, lim,
fi fi
if [ "$mode" = "lmstudio" ]; then if [ "$mode" = "lmstudio" ]; then
local cat; cat="$(lms_catalogue "$base")" local cat; cat="$(lms_catalogue "$base" "$(cm_preset_token "$pf")")"
if [ -n "$cat" ]; then if [ -n "$cat" ]; then
ok "LM Studio reachable at $base ($(printf '%s\n' "$cat" | wc -l | tr -d ' ') models installed)" ok "LM Studio reachable at $base ($(printf '%s\n' "$cat" | wc -l | tr -d ' ') models installed)"
local t id row st ctx local t id row st ctx
@@ -1617,10 +1662,10 @@ ui_pick_model() {
while IFS=$'\t' read -r id st ctx; do while IFS=$'\t' read -r id st ctx; do
[ -n "$id" ] || continue [ -n "$id" ] || continue
ids+=("$id"); ui_add_item "$id" "state: $st max context: $ctx" ids+=("$id"); ui_add_item "$id" "state: $st max context: $ctx"
done < <(lms_catalogue "$base") done < <(lms_catalogue "$base" "$(cm_preset_token "$pf")")
;; ;;
zai) zai)
ids+=('glm-5.2'); ui_add_item 'glm-5.2' 'flagship coding model - opus/sonnet tier' ids+=('glm-5.3'); ui_add_item 'glm-5.3' 'flagship coding model - opus/sonnet tier'
ids+=('glm-4.7'); ui_add_item 'glm-4.7' 'fast/cheap tier - haiku' ids+=('glm-4.7'); ui_add_item 'glm-4.7' 'fast/cheap tier - haiku'
;; ;;
esac esac
@@ -1764,6 +1809,468 @@ ui_menu() {
done done
} }
# ---------------------------------------------------------------------------
# First-run setup
#
# A shipped preset is a starting point, not a working configuration. OpenRouter
# needs a key before it can serve anything; LM Studio needs to be told where the
# server is and which of the models it actually has installed to use - and the
# ids it ships with are whatever happened to be on the machine this was written
# on, which is almost certainly not yours.
#
# So a preset says whether it has been through setup. `configured: false` is
# written into the shipped presets and cleared once setup has run, and preflight
# treats it as a blocker: better to be walked through it once than to switch
# into something that half-works and produces a confusing failure later.
#
# Absent means configured. That is deliberate - presets that predate this, and
# ones the user built by hand with `preset new`, are their own business and must
# not suddenly start demanding a wizard.
# ---------------------------------------------------------------------------
preset_configured() {
local v; v="$(jget "$1" configured)"
[ "$v" = "false" ] && return 1
return 0
}
mark_configured() {
"$PY" "$JSON" set-flag "$1" configured true >/dev/null 2>&1
}
# A y/N prompt that defaults to no on anything that is not a clear yes.
ask_yes() {
local prompt="$1" reply
printf ' %s [y/N] ' "$prompt"
IFS= read -r reply || return 1
case "$reply" in y|Y|yes|YES) return 0 ;; *) return 1 ;; esac
}
# Prompt with a default shown in brackets; empty input keeps the default.
#
# The prompt goes to stderr for the same reason warn/err do: this is called
# inside $( ), where anything on stdout is captured as the return value. Printed
# to stdout it came back as part of the answer - " server base URL [...]: " with
# the typed URL glued on the end, which set-url then rejected.
ask_value() {
local prompt="$1" default="$2" reply
if [ -n "$default" ]; then printf ' %s [%s]: ' "$prompt" "$default" >&2
else printf ' %s: ' "$prompt" >&2; fi
IFS= read -r reply || return 1
[ -n "$reply" ] && printf '%s' "$reply" || printf '%s' "$default"
}
setup_key() {
local ref="$1" label="$2"
if cm_vault_has "$ref"; then
ok "a key is already stored for '$ref' ($(cm_vault_backend_label))"
ask_yes "replace it?" || return 0
else
say "$label needs an API key. It goes into $(cm_vault_backend_label),"
say 'not into settings.json.'
fi
cmd_set_key "$ref"
}
# Offer the provider's own catalogue rather than asking someone to type a model
# id from memory. Falls back to typing when the catalogue cannot be reached,
# because being offline should not block finishing setup.
setup_models() {
local pf="$1" provider="$2" name="$3"
printf '\n'
say 'current model map:'
local t v
for t in "${TIERS[@]}"; do
v="$(jget "$pf" "models.$t")"
[ -n "$v" ] && printf ' %-8s %s\n' "$t" "$v"
done
printf '\n'
ask_yes 'change which models back these tiers?' || return 0
if ! ui_interactive; then
warn 'model picking needs an interactive terminal'
return 0
fi
if [ "$provider" = "lmstudio" ]; then
# One model for every tier is the normal shape for a local server: it
# has one loaded at a time, and mapping tiers to different models just
# means paying the load cost on every tier change.
local base ids=() id st ctx
base="$(jget "$pf" baseUrl)"
while IFS=$'\t' read -r id st ctx; do
[ -n "$id" ] || continue
ids+=("$id")
done < <(lms_catalogue "$base" "$(cm_preset_token "$pf")")
if [ "${#ids[@]}" -eq 0 ]; then
warn 'the server returned no models; type an id by hand instead'
local manual; manual="$(ask_value 'model id for every tier' "$(jget "$pf" models.opus)")"
[ -n "$manual" ] && "$PY" "$JSON" set-all "$pf" "$manual" >/dev/null && ok "all tiers -> $manual"
return 0
fi
ui_reset_items
for id in "${ids[@]}"; do ui_add_item "$id" 'use this for every tier'; done
if ui_filter_select "model for all tiers of '$name'" 'esc = keep current'; then
"$PY" "$JSON" set-all "$pf" "${ids[$UI_SEL]}" >/dev/null
ok "all tiers -> ${ids[$UI_SEL]}"
fi
return 0
fi
# Remote gateways map a different model per tier, which is the whole point
# of them, so each tier is asked for separately.
local cur
for t in "${TIERS[@]}"; do
cur="$(jget "$pf" "models.$t")"
ui_pick_model "$pf" "$t" "$cur" || continue
[ -n "$UI_PICKED" ] || continue
"$PY" "$JSON" set-tier "$pf" "$t" "$UI_PICKED" >/dev/null && ok "$t -> $UI_PICKED"
done
}
setup_lmstudio_server() {
local pf="$1" name="$2" url probe token
url="$(jget "$pf" baseUrl)"; [ -n "$url" ] || url='http://127.0.0.1:1234'
printf '\n'
say 'LM Studio does not have to be on this machine - a LAN address or'
say 'anything reachable through a tunnel or proxy works just as well.'
url="$(ask_value 'server base URL' "$url")"
"$PY" "$JSON" set-url "$pf" "$url" >/dev/null || return 1
ok "baseUrl -> $url"
printf '\n'
if ask_yes 'does that server require an API key?'; then
local ref; ref="$(ask_value 'key name to store it under' 'lmstudio')"
"$PY" "$JSON" set-auth "$pf" key "$ref" >/dev/null
ok "auth -> vault key '$ref'"
cm_vault_has "$ref" || cmd_set_key "$ref"
token="$(cm_vault_get "$ref" 2>/dev/null || true)"
else
"$PY" "$JSON" set-auth "$pf" none >/dev/null
ok 'auth -> none (inline placeholder token)'
token='lmstudio'
fi
printf '\n'
say "checking $url ..."
probe="$(cm_probe_server "$url" "$token")"
case "$probe" in
ok) ok 'server answered' ;;
auth) err 'the server refused that credential'; return 1 ;;
notfound) err 'something answered there, but not an LM Studio API' ; return 1 ;;
skip) warn 'curl is missing, so the server was not checked' ;;
*) err 'nothing answered at that address'
say 'start the server and run: claude-mode setup lmstudio'
return 1 ;;
esac
return 0
}
cmd_setup() {
local mode="${1:-}" name pf
case "$mode" in
anthropic)
head_ 'setup: anthropic'
ok 'nothing to configure - it uses your existing Claude login'
return 0 ;;
openrouter|zai|lmstudio) ;;
z.ai|z-ai) mode=zai ;;
'') err 'usage: claude-mode setup <mode>'; return 1 ;;
*) err "unknown mode '$mode'"; return 1 ;;
esac
name="$(resolve_preset "$mode" "${2:-}")" || return 1
pf="$(preset_path "$name")"
if ! ui_interactive; then
err 'setup needs an interactive terminal'
say "run: claude-mode setup $mode"
return 1
fi
head_ "setup: $mode / preset '$name'"
say "$(mode_label "$mode")"
case "$mode" in
openrouter)
printf '\n'
setup_key openrouter 'OpenRouter'
setup_models "$pf" openrouter "$name"
;;
zai)
printf '\n'
say 'get a key from https://z.ai/manage-apikey/apikey-list'
setup_key zai 'Z.AI'
setup_models "$pf" zai "$name"
;;
lmstudio)
setup_lmstudio_server "$pf" "$name" || return 1
setup_models "$pf" lmstudio "$name"
;;
esac
mark_configured "$pf"
printf '\n'
ok "$mode is set up"
say "switch to it with: claude-mode $mode"
return 0
}
# ---------------------------------------------------------------------------
# Live sessions: asked before the write, not reported after it
#
# The damage a switch does to a running session is not limited to it failing
# calls. If the session takes even one completion from the new provider before
# anything notices, that provider's message-id format lands in its transcript -
# OpenRouter issues `gen-<epoch>-<rand>` where Anthropic issues `msg_...` - and
# native Anthropic then refuses to resume the session at all:
#
# API Error: 400 diagnostics.previous_message_id: must be the `id` from a
# prior /v1/messages response (starts with `msg_`)
#
# There is no supported way back from that. The only fix is to truncate the
# transcript to the last message Anthropic issued, losing everything after it
# (see `claude-mode repair-session`). A confirmation that costs one keystroke is
# cheap against a failure that costs an afternoon of conversation.
# ---------------------------------------------------------------------------
CM_ASSUME_YES=0
CM_SESSION_ACTION=none
CM_SESSION_ROWS=''
cm_confirm_sessions() {
local mode="$1" rows n busy reply
CM_SESSION_ACTION=none
CM_SESSION_ROWS=''
rows="$(cm_session_rows)"
n="$(printf '%s' "$rows" | grep -c . || true)"
[ "${n:-0}" -gt 0 ] || return 0
CM_SESSION_ROWS="$rows"
busy="$(printf '%s' "$rows" | cut -f4 | grep -c '^yes$' || true)"
printf '\n'
warn "$n Claude Code session(s) are running right now"
local pid ppid tty b cwd isself pcmd tag
while IFS=$'\t' read -r pid ppid tty b cwd isself pcmd; do
[ -n "$pid" ] || continue
tag=''
[ "$b" = yes ] && tag=" ${C_YELLOW}working${C_RESET}"
[ "$isself" = yes ] && tag="$tag ${C_DIM}(this one)${C_RESET}"
printf ' %-8s %-8s %s%s\n' "$pid" "$tty" "$cwd" "$tag"
done <<EOF_ROWS
$rows
EOF_ROWS
printf '\n'
say 'Their key is re-fetched on a timer and will resolve to the new mode,'
say 'which the endpoint they are still pointed at will not accept. If one'
say 'of them takes a reply from the new provider first, that provider'"'"'s'
say 'message-id format goes into its transcript and Anthropic will then'
say 'refuse to resume that session at all - recoverable only by truncating'
say 'it (claude-mode repair-session), which loses the turns after the cut.'
[ "${busy:-0}" -gt 0 ] && warn "$busy of them is mid-request and is the most likely to be caught"
if [ "$CM_ASSUME_YES" -eq 1 ]; then
say 'proceeding (--yes)'
return 0
fi
if ! ui_interactive; then
printf '\n'
err 'refusing to switch while sessions are running'
say 'restart or close them first, or pass --yes to switch anyway'
return 1
fi
printf '\n'
say 'r switch, then close and reopen them on the new mode (safest)'
say 'c switch, then close them'
say 's switch and leave them running (risks the above)'
say 'a abort'
printf '\n [r/c/s/A] '
IFS= read -r reply
case "$reply" in
r|R) CM_SESSION_ACTION=restart; return 0 ;;
c|C) CM_SESSION_ACTION=stop; return 0 ;;
s|S) CM_SESSION_ACTION=none; return 0 ;;
*) say 'aborted; nothing was changed'; return 1 ;;
esac
}
# Run after the write, never before: a session reopened first would come back up
# on the mode being left behind.
cm_apply_session_action() {
[ "$CM_SESSION_ACTION" = "none" ] && return 0
[ -n "$CM_SESSION_ROWS" ] || return 0
printf '\n'
cm_session_act "$CM_SESSION_ACTION" "$CM_SESSION_ROWS" 0
}
# ---------------------------------------------------------------------------
# Transcript repair
# ---------------------------------------------------------------------------
# ~/.claude/projects/<cwd with every slash turned into a dash>
cm_project_dir() {
printf '%s/projects/%s' "$CM_SETTINGS_DIR" "$(printf '%s' "${1:-$PWD}" | sed 's|/|-|g')"
}
cmd_repair_session() {
local target='' apply=0 reinject=1 scan_all=0 as_json=0 dir='' a file verdict age
for a in "$@"; do
case "$a" in
--apply) apply=1 ;;
--dry-run) apply=0 ;;
--no-reinject) reinject=0 ;;
--list) target='--list' ;;
--all) scan_all=1 ;;
--json) as_json=1; scan_all=1 ;;
-*) err "unknown option '$a'"; return 1 ;;
*) target="$a" ;;
esac
done
# A session you need to repair is one you could not resume, which is a poor
# position from which to remember which project it belonged to. --all drops
# the working-directory scoping and reports only what is actually broken.
if [ "$scan_all" -eq 1 ]; then
local scan
scan="$("$PY" "$JSON" scan-sessions "$CM_SETTINGS_DIR/projects" 2>/dev/null)" || {
err 'could not scan transcripts'; return 1; }
if [ "$as_json" -eq 1 ]; then
printf '%s\n' "$scan"
return 0
fi
head_ 'scanning every session transcript'
printf '%s' "$scan" | "$PY" -c "
import json,sys
d=json.load(sys.stdin)
G,Y,D,X='\033[32m','\033[33m','\033[90m','\033[0m'
for b in d['broken']:
print(' %-38s %s' % (b['sessionId'], b['project']))
prov = ', '.join(b['providers']) or 'another provider'
print(' %d line(s) after the last good message, from %s' % (b['dropLines'], prov))
print()
if not d['broken']:
print(' %sok %s nothing to repair across %d transcript(s)' % (G,X,d['scanned']))
else:
print(' %swarn%s %d of %d transcript(s) were cut short by a mode switch' % (Y,X,d['count'],d['scanned']))
print(' %sclaude-mode repair-session <id> --apply%s' % (D,X))
print(' %ssessions that ran entirely on a gateway are not listed: they carry that%s' % (D,X))
print(' %sprovider\'s ids by design and resume fine under it%s' % (D,X))
"
return 0
fi
# Claude Code keys transcripts by the directory the session was started in,
# which is rarely the one you are standing in when you come to fix it. Walk
# up first, and for a named session fall back to looking through every
# project - the id is unique, so there is nothing ambiguous to resolve.
local probe="$PWD"
while [ -n "$probe" ]; do
[ -d "$(cm_project_dir "$probe")" ] && { dir="$(cm_project_dir "$probe")"; break; }
[ "$probe" = "/" ] && break
probe="$(dirname "$probe")"
done
if [ -n "$target" ] && [ "$target" != "--list" ]; then
if [ -z "$dir" ] || [ ! -f "$dir/${target%.jsonl}.jsonl" ]; then
local hit
hit="$(ls -1 "$CM_SETTINGS_DIR"/projects/*/"${target%.jsonl}".jsonl 2>/dev/null | head -n1)"
[ -n "$hit" ] && dir="$(dirname "$hit")"
fi
fi
if [ -z "$dir" ] || [ ! -d "$dir" ]; then
err 'no session transcripts found for this directory'
say 'run it from the project the session belongs to, or name the session id'
return 1
fi
if [ -z "$target" ] || [ "$target" = "--list" ]; then
head_ 'session transcripts here'
local f v
for f in $(ls -1t "$dir"/*.jsonl 2>/dev/null); do
v="$("$PY" "$JSON" repair-session "$f" 2>/dev/null)" || continue
printf '%s' "$v" | "$PY" -c "
import json,sys,os
d=json.load(sys.stdin)
state = 'ok' if d['healthy'] else ('repairable, would drop %d line(s)' % d['dropLines'] if d['repairable'] else 'no anthropic message found')
print(' %-40s %s' % (os.path.basename(d['path'])[:-6], state))
"
done
printf '\n %sclaude-mode repair-session <session-id> --apply%s\n' "$C_DIM" "$C_RESET"
printf ' %s--all checks every project, not just this one%s\n' "$C_DIM" "$C_RESET"
return 0
fi
file="$dir/${target%.jsonl}.jsonl"
[ -f "$file" ] || { err "no transcript $file"; return 1; }
# A transcript that is still being appended to belongs to a session that is
# still alive; truncating it underneath a running process helps nobody.
age=$(( $(date +%s) - $(stat -c %Y "$file" 2>/dev/null || echo 0) ))
if [ "$age" -lt 90 ] && [ "$apply" -eq 1 ]; then
err "that transcript was written to ${age}s ago - it looks live"
say 'close the session that owns it first'
return 1
fi
local flags=''
[ "$apply" -eq 1 ] && flags="$flags --apply"
[ "$reinject" -eq 0 ] && flags="$flags --no-reinject"
# shellcheck disable=SC2086
verdict="$("$PY" "$JSON" repair-session "$file" $flags)" || {
err 'could not read that transcript'; return 1; }
printf '%s' "$verdict" | "$PY" -c "
import json,sys
d=json.load(sys.stdin)
G,Y,R,D,X = '\033[32m','\033[33m','\033[91m','\033[90m','\033[0m'
print()
if d['healthy']:
print(' %sok %s last message is Anthropic-issued; nothing to repair' % (G,X))
raise SystemExit(0)
if d['kind'] == 'gateway-native':
prov = (d['foreignIds'][0]['model'] or 'a gateway') if d['foreignIds'] else 'a gateway'
print(' %sok %s this session ran entirely on %s' % (G,X,prov))
print(' %sits ids come from that provider by design; it resumes under it, not%s' % (D,X))
print(' %sunder Anthropic. There is nothing here to repair.%s' % (D,X))
raise SystemExit(0)
if not d['repairable']:
print(' %sok %s this transcript has no assistant replies to resume from' % (G,X))
raise SystemExit(0)
for f in d['foreignIds']:
print(' %swarn%s line %d carries a %s id from %s' % (Y,X,f['line'],f['id'].split('-')[0]+'-',f['model'] or 'another provider'))
n = sum(1 for s in d['syntheticIds'] if s['apiError'] and s['line'] > d['lastGoodLine'])
if n:
print(' %swarn%s %d client-side error placeholder(s) after the last good message' % (Y,X,n))
if d['applied']:
print(' %sok %s truncated to line %d, dropping %d' % (G,X,d['lastGoodLine'],d['dropLines']))
print(' %sok %s original saved as %s' % (G,X,d['backup']))
if d.get('recovered'):
print(' %sok %s dropped turns written to %s' % (G,X,d['recovered']))
if d.get('reinjected'):
print(' %sok %s and handed back to the session as a context note' % (G,X))
print()
print(' %sthat session should resume, and will know what it did%s' % (D,X))
else:
print(' %swarn%s would truncate to line %d, dropping %d line(s)' % (Y,X,d['lastGoodLine'],d['dropLines']))
print()
print(' %sre-run with --apply: the original is backed up, the dropped turns are%s' % (D,X))
print(' %ssaved as markdown, and handed back to the session as a context note%s' % (D,X))
print(' %s(--no-reinject writes the file but leaves the session untouched)%s' % (D,X))
"
}
# --------------------------------------------------------------------------- # ---------------------------------------------------------------------------
# Dispatch # Dispatch
# --------------------------------------------------------------------------- # ---------------------------------------------------------------------------
@@ -1777,6 +2284,7 @@ _args=()
for _a in "$@"; do for _a in "$@"; do
case "$_a" in case "$_a" in
--force) CM_FORCE=1 ;; --force) CM_FORCE=1 ;;
-y|--yes) CM_ASSUME_YES=1 ;;
*) _args+=("$_a") ;; *) _args+=("$_a") ;;
esac esac
done done
@@ -1817,7 +2325,29 @@ case "$cmd" in
doctor) cmd_doctor ;; doctor) cmd_doctor ;;
health) write_health "$(state_mode)" "$(state_preset)" ;; health) write_health "$(state_mode)" "$(state_preset)" ;;
preflight) cmd_preflight "${1:-}" "${2:-}" ;; preflight) cmd_preflight "${1:-}" "${2:-}" ;;
setup)
# --terminal for callers with no stdin to offer. The bar widget
# cannot host a hidden key prompt or a filter-select list, so it
# asks a terminal to host the whole flow instead.
_mode="${1:-}"; _term=0; _rest=''
for _a in "$@"; do
case "$_a" in
--terminal) _term=1 ;;
"$_mode") ;;
*) _rest="$_a" ;;
esac
done
if [ "$_term" -eq 1 ]; then
_t="$(cm_terminal_cmd)"
setsid nohup "$_t" -e bash -lc \
"'$0' setup '$_mode' $_rest; printf '\n press enter to close '; read -r _" \
>/dev/null 2>&1 &
ok "opened $_t to set up $_mode"
else
cmd_setup "$_mode" "$_rest"
fi ;;
sessions) cmd_sessions "$@" ;; sessions) cmd_sessions "$@" ;;
repair-session) cmd_repair_session "$@" ;;
repair) repair)
scope='' scope=''
for a in "$@"; do [ "$a" = "--all" ] && scope=all; done for a in "$@"; do [ "$a" = "--all" ] && scope=all; done
linux/cm-json.py
+367
View File
@@ -16,6 +16,7 @@ Subcommands:
presets <dir> "name<TAB>provider<TAB>desc" presets <dir> "name<TAB>provider<TAB>desc"
""" """
import glob
import json import json
import os import os
import re import re
@@ -579,10 +580,374 @@ def cmd_set_auth(argv):
save(path, p) save(path, p)
print(json.dumps(p["auth"])) print(json.dumps(p["auth"]))
def cmd_set_flag(argv):
"""set-flag <preset> <key> true|false
Only used for `configured` so far. Kept generic because a preset-level
boolean written by hand is exactly the kind of thing that ends up as a
string "false", which is truthy everywhere that matters.
"""
path, key, val = argv[0], argv[1], argv[2]
p = load(path)
p[key] = (val == "true")
save(path, p)
def cmd_set_all(argv):
"""set-all <preset> <model-id> - point every tier and the subagent at one id."""
path, model = argv[0], argv[1]
p = load(path)
p["models"] = {t: model for t in TIERS}
if p.get("subagentModel") and p["subagentModel"] != "inherit":
p["subagentModel"] = model
save(path, p)
print(model)
# ---------------------------------------------------------------------------
# Session transcript repair
#
# Native Anthropic requires previous_message_id to be an id it issued itself -
# one starting `msg_`. A session that took even one completion from a gateway
# while the mode was switched under it has that provider's id format in its
# transcript instead (OpenRouter issues `gen-<epoch>-<rand>`), and every attempt
# to resume it afterwards fails with a 400 naming previous_message_id. Client-
# side error placeholders, written as model `<synthetic>` with a UUID for an id,
# do the same thing when one is last.
#
# The transcript is newline-delimited JSON, one independent object per line, so
# rolling back to the last message Anthropic actually issued is a truncation.
# Everything after it is lost - which is the cost, and why nothing here runs
# without being asked twice.
# ---------------------------------------------------------------------------
def _msg_id(obj):
m = obj.get("message")
if isinstance(m, dict) and m.get("id"):
return str(m["id"])
return None
def _blocks(msg):
c = msg.get("content")
if isinstance(c, str):
return [{"type": "text", "text": c}]
return c if isinstance(c, list) else []
def _clip(t, n):
t = " ".join(str(t).split())
return t if len(t) <= n else t[:n - 1] + "\u2026"
def summarise_dropped(lines, per_block=1400):
"""Render the turns a repair is about to discard as readable Markdown.
Tool *results* are deliberately left out. They are the bulk of a transcript
by volume and the least useful part of a summary - what matters on the way
back in is what was asked, what was said, and what was run.
"""
out, tools = [], 0
for line in lines:
if not line.strip():
continue
try:
d = json.loads(line)
except Exception:
continue
t, msg = d.get("type"), (d.get("message") or {})
if t == "user" and not d.get("isMeta"):
for b in _blocks(msg):
if b.get("type") == "text" and b.get("text", "").strip():
out.append("### User\n\n" + _clip(b["text"], per_block))
elif t == "assistant":
if d.get("isApiErrorMessage"):
for b in _blocks(msg):
if b.get("type") == "text":
out.append("> **API error:** " + _clip(b.get("text", ""), 300))
continue
said, ran = [], []
for b in _blocks(msg):
if b.get("type") == "text" and b.get("text", "").strip():
said.append(_clip(b["text"], per_block))
elif b.get("type") == "tool_use":
inp = b.get("input") or {}
hint = inp.get("command") or inp.get("file_path") or inp.get("pattern") or inp.get("path") or ""
ran.append("`%s`%s" % (b.get("name", "tool"),
(" \u2014 " + _clip(hint, 120)) if hint else ""))
tools += 1
if said:
out.append("### Claude\n\n" + "\n\n".join(said))
if ran:
out.append("Ran: " + ", ".join(ran[:12]) + (" \u2026" if len(ran) > 12 else ""))
return "\n\n".join(out), tools
def _meta_entry(template, parent_uuid, text):
"""A user-role entry marked isMeta, which Claude Code treats as context
rather than as something to answer - the same marker it uses for its own
local-command caveats. It carries no message.id, so it cannot affect the
previous_message_id that made the session unresumable in the first place.
"""
now = __import__("datetime").datetime.now(
__import__("datetime").timezone.utc).strftime("%Y-%m-%dT%H:%M:%S.000Z")
e = {
"parentUuid": parent_uuid,
"isSidechain": False,
"userType": template.get("userType", "external"),
"cwd": template.get("cwd", ""),
"sessionId": template.get("sessionId", ""),
"version": template.get("version", ""),
"gitBranch": template.get("gitBranch", ""),
"type": "user",
"isMeta": True,
"uuid": str(__import__("uuid").uuid4()),
"timestamp": now,
"message": {"role": "user", "content": text},
}
return {k: v for k, v in e.items() if v != "" or k in ("gitBranch",)}
def cmd_repair_session(argv):
"""repair-session <transcript.jsonl> [--apply] [--no-reinject]
Prints a JSON verdict. With --apply, writes the dropped turns out as
Markdown, backs the file up, truncates it to the last Anthropic-issued
message, and (unless --no-reinject) appends those turns back as a single
isMeta note so the resumed session still knows what it did.
"""
path = argv[0]
apply_it = "--apply" in argv[1:]
reinject = "--no-reinject" not in argv[1:]
with open(path, encoding="utf-8") as fh:
lines = fh.read().splitlines()
last_good = -1 # index of the last line carrying a msg_ id
foreign, synthetic = [], []
for i, line in enumerate(lines):
if not line.strip():
continue
try:
obj = json.loads(line)
except Exception:
continue
mid = _msg_id(obj)
if mid is None:
continue
if mid.startswith("msg_"):
last_good = i
elif mid.startswith("gen-"):
foreign.append({"line": i + 1, "id": mid,
"model": str((obj.get("message") or {}).get("model", ""))})
else:
synthetic.append({"line": i + 1, "id": mid,
"apiError": bool(obj.get("isApiErrorMessage"))})
# Only the *last* id matters for resuming: an error placeholder in the
# middle of a long-finished turn is history, not a blocker.
tail_ids = [i for i in range(len(lines) - 1, last_good, -1)
if lines[i].strip() and _msg_id(_safe(lines[i])) is not None]
healthy = (last_good >= 0 and not tail_ids)
# Not every transcript without an Anthropic id is damaged, and saying so
# turns the scan into noise. A session that ran start to finish on a gateway
# has `gen-` ids throughout by design: it resumes perfectly well under the
# provider it was born on, has nothing to truncate back to, and is only a
# problem if you try to resume it as Anthropic. Likewise a session that never
# got a reply at all has no ids and nothing wrong with it.
if healthy:
kind = "healthy"
elif last_good >= 0:
kind = "repairable"
elif foreign:
kind = "gateway-native"
elif synthetic:
kind = "synthetic-only"
else:
kind = "no-messages"
out = {
"path": path,
"lines": len(lines),
"lastGoodLine": last_good + 1 if last_good >= 0 else 0,
"dropLines": 0 if healthy or last_good < 0 else len(lines) - (last_good + 1),
"foreignIds": foreign,
"syntheticIds": synthetic,
"healthy": healthy,
"kind": kind,
"repairable": (not healthy) and last_good >= 0,
"applied": False,
"backup": "",
"recovered": "",
"reinjected": False,
"droppedTools": 0,
}
if apply_it and out["repairable"]:
stamp = __import__("datetime").datetime.now().strftime("%Y%m%d-%H%M%S")
backup = "%s.pre-repair-backup-%s" % (path, stamp)
dropped = lines[last_good + 1:]
keep = lines[:last_good + 1]
# The turns being cut are the work itself. Losing the conversation that
# produced a morning's changes is most of the damage, so they are written
# out as readable Markdown before anything is deleted...
digest, ntools = summarise_dropped(dropped)
recovered = "%s.recovered-%s.md" % (path.rsplit(".jsonl", 1)[0], stamp)
header = ("# Recovered turns\n\n"
"Cut from `%s` on %s, because the transcript could no longer be\n"
"resumed. Everything below happened; none of it is in the session any more.\n\n"
"---\n\n" % (os.path.basename(path), stamp))
with open(recovered, "w", encoding="utf-8") as fh:
fh.write(header + (digest or "_No readable messages in the dropped turns._\n"))
with open(backup, "w", encoding="utf-8") as fh:
fh.write("\n".join(lines) + ("\n" if lines else ""))
# ...and then handed back to the session, so the agent that resumes it
# knows what it just did rather than waking with a gap in its memory.
if reinject and digest:
template = _safe(keep[last_good])
body = digest if len(digest) <= 24000 else (
digest[:24000] + "\n\n\u2026 truncated; the full text is in " + recovered)
note = ("<recovered-transcript>\n"
"Context, not an instruction - do not act on it or reply to it.\n\n"
"This session was rolled back to its last resumable point after a "
"provider switch left it unable to resume. The turns below were part "
"of this conversation and are no longer in it. The full text is at "
+ recovered + "\n\n" + body + "\n</recovered-transcript>")
keep.append(json.dumps(_meta_entry(template, template.get("uuid"), note)))
out["reinjected"] = True
with open(path, "w", encoding="utf-8") as fh:
fh.write("\n".join(keep) + "\n")
out["applied"] = True
out["backup"] = backup
out["recovered"] = recovered
out["droppedTools"] = ntools
print(json.dumps(out))
def _safe(line):
try:
return json.loads(line)
except Exception:
return {}
def _last_msg_id_from_tail(path, window=262144):
"""The id of the last message-bearing line, read from the end of the file.
Classifying a transcript needs two facts: what the last message id is, and
whether any Anthropic id exists at all. The first decides the common case on
its own - if the last one is `msg_`, the transcript is healthy and the rest
of the file never has to be touched. Since that is 40-odd of the 59 here,
and some are 10MB, reading the tail first is the difference between a scan
that can run on a timer and one that cannot.
Returns (id_or_None, saw_whole_file).
"""
size = os.path.getsize(path)
with open(path, "rb") as fh:
start = max(0, size - window)
fh.seek(start)
chunk = fh.read()
whole = start == 0
if not whole:
nl = chunk.find(b"\n") # drop the partial first line
chunk = chunk[nl + 1:] if nl >= 0 else b""
for raw in reversed(chunk.splitlines()):
if not raw.strip():
continue
try:
d = json.loads(raw.decode("utf-8", "replace"))
except Exception:
continue
mid = _msg_id(d)
if mid:
return mid, whole
return None, whole
def cmd_scan_sessions(argv):
"""scan-sessions <projects-dir> - JSON list of transcripts worth repairing.
Only `repairable` is reported: a transcript carrying a genuine Anthropic
message with a different provider's output after it. A session that ran
entirely on a gateway, or never got a reply, is not damage.
"""
root = argv[0]
broken, scanned, skipped = [], 0, 0
for proj in sorted(glob.glob(os.path.join(root, "*"))):
if not os.path.isdir(proj):
continue
for path in sorted(glob.glob(os.path.join(proj, "*.jsonl"))):
scanned += 1
try:
last, whole = _last_msg_id_from_tail(path)
except OSError:
skipped += 1
continue
# Healthy is decided by the tail alone, and is the common case.
if last is not None and last.startswith("msg_"):
continue
if last is None and whole:
continue # no replies at all; nothing to fix
try:
with open(path, encoding="utf-8", errors="replace") as fh:
lines = fh.read().splitlines()
except OSError:
skipped += 1
continue
last_good, foreign = -1, []
for i, line in enumerate(lines):
if not line.strip():
continue
try:
d = json.loads(line)
except Exception:
continue
mid = _msg_id(d)
if not mid:
continue
if mid.startswith("msg_"):
last_good = i
elif mid.startswith("gen-"):
foreign.append(str((d.get("message") or {}).get("model", "")))
if last_good < 0:
continue # gateway-native or reply-less
tail = [i for i in range(len(lines) - 1, last_good, -1)
if lines[i].strip() and _msg_id(_safe(lines[i])) is not None]
if not tail:
continue
broken.append({
"sessionId": os.path.basename(path)[:-6],
"project": os.path.basename(proj),
"path": path,
"lines": len(lines),
"dropLines": len(lines) - (last_good + 1),
"providers": sorted(set(f for f in foreign if f)),
"mtime": int(os.path.getmtime(path)),
})
print(json.dumps({"scanned": scanned, "skipped": skipped,
"broken": broken, "count": len(broken)}))
COMMANDS = { COMMANDS = {
"health": cmd_health, "health": cmd_health,
"preflight-json": cmd_preflight_json, "preflight-json": cmd_preflight_json,
"sessions-json": cmd_sessions_json, "sessions-json": cmd_sessions_json,
"repair-session": cmd_repair_session,
"scan-sessions": cmd_scan_sessions,
"stale-models": cmd_stale_models, "stale-models": cmd_stale_models,
"strip-tags": cmd_strip_tags, "strip-tags": cmd_strip_tags,
"or-models": cmd_or_models, "or-models": cmd_or_models,
@@ -592,6 +957,8 @@ COMMANDS = {
"models": cmd_models, "models": cmd_models,
"set-tier": cmd_set_tier, "set-tier": cmd_set_tier,
"set-url": cmd_set_url, "set-url": cmd_set_url,
"set-flag": cmd_set_flag,
"set-all": cmd_set_all,
"set-auth": cmd_set_auth, "set-auth": cmd_set_auth,
"scaffold": cmd_scaffold, "scaffold": cmd_scaffold,
"get": cmd_get, "get": cmd_get,
omarchy/install.sh
Executable
+91
View File
@@ -0,0 +1,91 @@
#!/usr/bin/env bash
# Installs the claude-mode bar widget into the Omarchy shell.
#
# Two steps: drop the plugin into ~/.config/omarchy/plugins/, and add its id to
# the bar layout in ~/.config/omarchy/shell.json. Both are idempotent, and the
# shell hot-reloads each of them, so nothing has to be restarted.
#
# Only ~/.config is touched. /usr/share/omarchy is owned by the package and is
# rewritten by `omarchy update`, so nothing may be installed there.
set -euo pipefail
PLUGIN_ID="smoido.claude-mode"
SRC="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)/$PLUGIN_ID"
DEST_DIR="$HOME/.config/omarchy/plugins/$PLUGIN_ID"
SHELL_JSON="$HOME/.config/omarchy/shell.json"
# Where in the bar the widget lands, and which existing widget it sits before.
# The AI-adjacent group on the right is the natural neighbourhood for it.
SECTION="${CM_BAR_SECTION:-right}"
BEFORE="${CM_BAR_BEFORE:-omarchy.agents}"
green() { printf ' \033[32mok \033[0m %s\n' "$*"; }
warn() { printf ' \033[33mwarn\033[0m %s\n' "$*"; }
fail() { printf ' \033[31mFAIL\033[0m %s\n' "$*"; }
printf '\ninstalling %s -> %s\n' "$PLUGIN_ID" "$DEST_DIR"
[ -d "$SRC" ] || { fail "plugin source not found: $SRC"; exit 1; }
command -v python3 >/dev/null 2>&1 || { fail 'python3 required'; exit 1; }
if [ ! -d "$HOME/.config/omarchy" ]; then
fail 'no ~/.config/omarchy - this does not look like an Omarchy system'
exit 1
fi
# --- payload ---------------------------------------------------------------
mkdir -p "$DEST_DIR"
for f in manifest.json BarWidget.qml Panel.qml BrandIcon.qml Modes.js; do
[ -f "$SRC/$f" ] || { fail "missing from payload: $f"; exit 1; }
install -m 0644 "$SRC/$f" "$DEST_DIR/$f"
done
green 'plugin files copied'
# --- bar layout ------------------------------------------------------------
# shell.json is the user's own file and may carry unrelated customisation, so
# it is read, minimally amended, and written back rather than templated over.
if [ ! -f "$SHELL_JSON" ]; then
warn "no $SHELL_JSON - add the widget yourself once the shell writes one"
exit 0
fi
cp "$SHELL_JSON" "$SHELL_JSON.bak.$(date +%s)"
python3 - "$SHELL_JSON" "$PLUGIN_ID" "$SECTION" "$BEFORE" <<'PYEOF'
import json, sys
path, plugin_id, section, before = sys.argv[1:5]
with open(path, encoding="utf-8") as fh:
cfg = json.load(fh)
layout = cfg.setdefault("bar", {}).setdefault("layout", {})
entries = layout.setdefault(section, [])
def has(entry_list):
return any(isinstance(e, dict) and e.get("id") == plugin_id for e in entry_list)
# Already placed anywhere in the bar: leave the user's chosen position alone.
for name, items in layout.items():
if isinstance(items, list) and has(items):
print(" already in bar layout (%s) - position left as-is" % name)
raise SystemExit(0)
idx = len(entries)
for i, e in enumerate(entries):
if isinstance(e, dict) and e.get("id") == before:
idx = i
break
entries.insert(idx, {"id": plugin_id})
with open(path, "w", encoding="utf-8") as fh:
json.dump(cfg, fh, indent=2)
fh.write("\n")
print(" added to bar.layout.%s at position %d" % (section, idx))
PYEOF
green 'shell.json updated (hot-reloads; no restart needed)'
printf '\ndone. The icon shows the active mode; click it to switch.\n'
printf 'If it does not appear: omarchy-shell shell rescanPlugins\n'
omarchy/smoido.claude-mode/BarWidget.qml
+330
View File
@@ -0,0 +1,330 @@
import QtQuick
import Quickshell.Io
import qs.Commons
import qs.Ui
import "Modes.js" as Modes
// Bar slot for claude-mode: shows which provider the next `claude` launch will
// use, and opens a panel that switches it.
//
// State comes from ~/.claude-mode/health.json, which claude-mode rewrites on
// every switch and on every `status`. Watching that file rather than polling
// the CLI keeps this widget at zero cost while idle, and means a switch made
// in a terminal shows up here without anything having to tell us.
BarWidget {
id: root
moduleName: "smoido.claude-mode"
readonly property string cmRoot: String(setting("root", Modes.defaultRoot()))
// Off by default - the bar is scarce horizontal space and the mark already
// says which provider is active. Turn it on per-instance in shell.json to get
// the preset name back beside the icon.
readonly property bool showLabel: setting("showLabel", false) === true
// Whole health.json, or null before the first successful read.
property var health: null
// state.json is the narrower fallback: it exists from install onward, where
// health.json only appears once claude-mode has run at least once. Without
// this the widget would render "unknown" on a fresh install until the first
// switch, which is precisely when someone is most likely to look at it.
property var fallbackState: null
readonly property string mode: {
if (health && health.mode) return String(health.mode)
if (fallbackState && fallbackState.mode) return String(fallbackState.mode)
return ""
}
readonly property string preset: {
if (mode === "anthropic") return ""
if (health && health.preset) return String(health.preset)
if (fallbackState && fallbackState.preset) return String(fallbackState.preset)
return ""
}
readonly property bool known: mode !== ""
// Measured against the neighbours: every stock bar glyph in this shell paints
// 11px of ink from a 13px font. These marks fill their box rather than
// carrying a font's internal padding, so the box itself has to be the smaller
// number or the widget sits visibly larger than everything beside it.
// Odd on purpose. The marks are radially symmetric, so their vertical and
// horizontal arms sit on the centre line - which lands on a pixel *centre* at
// an odd size and on the boundary between two pixels at an even one, where
// each arm splits its coverage and comes out grey on both sides.
readonly property int iconPx: {
var n = Math.round(Number(root.setting("iconSize", Style.bar.iconFont)))
return n % 2 === 0 ? n + 1 : n
}
readonly property string logo: Modes.logo(mode)
readonly property string glyph: Modes.glyph(mode)
readonly property string label: Modes.shortLabel(mode, preset)
// A gateway mode is spending money or leaning on a local server; native
// Anthropic is the resting state. Only the former earns the accent, so the
// bar stays quiet exactly when nothing unusual is configured.
readonly property bool gateway: known && mode !== "anthropic"
readonly property color urgentColor: {
var c = Color.urgent
var mx = Math.max(c.r, c.g, c.b)
var mn = Math.min(c.r, c.g, c.b)
var l = (mx + mn) / 2
var d = 1 - Math.abs(2 * l - 1)
return (d > 0.0001 ? (mx - mn) / d : 0) >= 0.15 ? c : "#d2685f"
}
readonly property color activeColor: !known
? Color.muted
: (gateway ? Color.accent : (bar ? bar.barForeground : Color.foreground))
// The bar's shared tooltip only paints for a target that reports itself
// hovered through this exact property.
readonly property bool tooltipHovered: visible && opacity > 0 && hoverHandler.hovered
function parseInto(prop, content) {
try {
var parsed = JSON.parse(String(content || ""))
root[prop] = (parsed && typeof parsed === "object") ? parsed : null
} catch (e) {
root[prop] = null
}
}
// `text()` is stale inside onFileChanged, so both first load and every later
// change are routed through reload() -> onLoaded to always parse fresh bytes.
FileView {
id: healthFile
path: root.cmRoot + "/health.json"
watchChanges: true
printErrors: false
onFileChanged: reload()
onLoaded: root.parseInto("health", text())
onLoadFailed: root.health = null
}
FileView {
id: stateFile
path: root.cmRoot + "/state.json"
watchChanges: true
printErrors: false
onFileChanged: reload()
onLoaded: root.parseInto("fallbackState", text())
onLoadFailed: root.fallbackState = null
}
// claude-mode writes health.json itself, so a switch launched from the panel
// lands back here through the FileView above. This only covers the case
// where the file was never written at all.
Process {
id: seedProc
command: [root.cmRoot + "/bin/claude-mode", "health"]
running: false
}
// ---- Broken transcripts
//
// A session cut short by a mode switch cannot be resumed, and nothing tells
// you until you try - by which time you have usually forgotten which session
// it was. So the bar checks periodically and marks itself when there is
// something to fix.
//
// The scan reads the tail of each transcript first and only opens the whole
// file when the tail already looks wrong, which is what makes it cheap enough
// to sit on a timer: ~60ms for 59 transcripts here, against 5s for the naive
// version that read every byte of every one.
property var brokenSessions: []
Process {
id: scanProc
running: false
command: [root.cmRoot + "/bin/claude-mode", "repair-session", "--json"]
stdout: StdioCollector {
waitForEnd: true
onStreamFinished: {
var d = null
try { d = JSON.parse(String(text || "")) } catch (e) { d = null }
root.brokenSessions = (d && d.broken) ? d.broken : []
}
}
}
function scanSessions() { if (!scanProc.running) scanProc.running = true }
Timer {
interval: 20 * 60 * 1000
running: true
repeat: true
triggeredOnStart: true
onTriggered: root.scanSessions()
}
function refresh() {
healthFile.reload()
stateFile.reload()
if (!root.health) seedProc.running = true
root.scanSessions()
}
Component.onCompleted: Qt.callLater(root.refresh)
// ---- Bar content
readonly property int gap: Style.space(5)
readonly property real iconWidth: root.logo !== "" ? icon.implicitWidth : fallbackIcon.implicitWidth
implicitWidth: vertical
? barSize
: (iconWidth + (labelText.visible ? gap + labelText.implicitWidth : 0) + Style.space(10))
implicitHeight: vertical ? (iconPx + Style.space(10)) : barSize
Row {
anchors.centerIn: parent
spacing: root.gap
Item {
anchors.verticalCenter: parent.verticalCenter
width: icon.width
height: icon.height
visible: root.logo !== ""
BrandIcon {
id: icon
anchors.fill: parent
pathData: root.logo
opticalScale: Modes.logoScale(root.mode)
color: root.activeColor
iconSize: root.iconPx
}
// Recolouring the mark would misreport the mode, which is this widget's
// whole job, so the warning gets its own dot instead.
Rectangle {
visible: root.brokenSessions.length > 0
width: Math.max(4, Math.round(root.iconPx / 3.2))
height: width
radius: width / 2
color: root.urgentColor
anchors.right: parent.right
anchors.top: parent.top
anchors.rightMargin: -Math.round(width / 3)
anchors.topMargin: -Math.round(width / 4)
}
}
// Only reached when claude-mode is not installed - there is no brand to
// draw for "no answer", so the pictograph stands in.
Text {
id: fallbackIcon
anchors.verticalCenter: parent.verticalCenter
visible: root.logo === ""
text: root.glyph
color: root.activeColor
font.family: root.bar ? root.bar.fontFamily : Style.font.family
font.pixelSize: Style.bar.iconFont
renderType: Text.NativeRendering
}
Text {
id: labelText
anchors.verticalCenter: parent.verticalCenter
visible: !root.vertical && root.showLabel && root.label !== ""
text: root.label
color: root.activeColor
font.family: root.bar ? root.bar.fontFamily : Style.font.family
font.pixelSize: Style.bar.iconFont - 1
renderType: Text.NativeRendering
}
}
// ---- Interaction
// The bar decides a module is clickable - and gives it the pointing-hand
// cursor - purely by whether it exposes triggerPress().
function triggerPress(button) {
if (root.bar) root.bar.hideTooltip(root)
root.togglePanel()
}
function tooltipLabel() {
if (!root.known) return "claude-mode: not installed"
var lines = [Modes.title(root.mode)]
if (root.preset !== "") lines.push("preset " + root.preset)
var models = root.health && root.health.models ? root.health.models : null
if (models && models.opus) lines.push("opus " + models.opus)
if (models && models.sonnet) lines.push("sonnet " + models.sonnet)
if (root.brokenSessions.length > 0) {
lines.push(root.brokenSessions.length === 1
? "1 session needs repair"
: root.brokenSessions.length + " sessions need repair")
}
return lines.join(" · ")
}
function syncTooltip() {
if (!bar || !hoverHandler.hovered) return
var text = tooltipLabel()
if (bar.tooltipTarget === root) {
if (bar.tooltipText !== text) bar.tooltipText = text
} else {
bar.showTooltip(root, text)
}
}
onHealthChanged: syncTooltip()
onBrokenSessionsChanged: syncTooltip()
// Hover must come from a HoverHandler: once triggerPress() exists the bar's
// own slot MouseArea accepts hover events and swallows them before any
// MouseArea here would see them. The call is deferred one turn because
// `hovered` flips before tooltipHovered has propagated, and showTooltip()
// silently refuses a target that does not yet report itself hovered.
HoverHandler {
id: hoverHandler
onHoveredChanged: {
if (!root.bar) return
if (hovered) Qt.callLater(root.syncTooltip)
else root.bar.hideTooltip(root)
}
}
MouseArea {
anchors.fill: parent
acceptedButtons: Qt.RightButton | Qt.MiddleButton
onClicked: function(mouse) {
if (mouse.button === Qt.MiddleButton) root.refresh()
else if (panelLoader.item) panelLoader.item.switchTo("anthropic", "")
}
}
// ---- Panel wiring
function injectPanel() {
var target = panelLoader.item
if (!target) return
if ("bar" in target) target.bar = root.bar
if ("settings" in target) target.settings = root.settings
if ("anchorItem" in target) target.anchorItem = root
if ("hostWidget" in target) target.hostWidget = root
if ("widget" in target) target.widget = root
}
function togglePanel() { if (panelLoader.item) panelLoader.item.toggle() }
readonly property bool opened: panelLoader.item ? panelLoader.item.opened === true : false
function open() { if (panelLoader.item) panelLoader.item.open() }
function close() { if (panelLoader.item) panelLoader.item.close() }
onBarChanged: injectPanel()
onSettingsChanged: injectPanel()
Loader {
id: panelLoader
active: true
source: Qt.resolvedUrl("Panel.qml")
visible: false
onLoaded: {
root.injectPanel()
Qt.callLater(root.injectPanel)
}
}
}
omarchy/smoido.claude-mode/BrandIcon.qml
+62
View File
@@ -0,0 +1,62 @@
import QtQuick
import QtQuick.Shapes
import qs.Commons
// One provider logo, drawn as a vector path and filled with whatever colour it
// is handed - which is what lets these follow the theme the way a font glyph
// would, instead of being a fixed-colour bitmap that looks pasted on.
//
// Every mark is authored in a 24x24 viewBox (see Modes.js) and scaled to the
// requested size from there.
//
// Two things about how that scaling is done, both of which matter at bar sizes:
//
// 1. No `layer.enabled`. A layer renders the Shape into a texture at the item's
// own size and then scales the *texture*, so a 24x24 buffer minified to 16px
// resamples 1.5 pixels into 1. On the Claude burst, whose rays are under a
// pixel wide at that size, that is not softening - rays merge, drop out, and
// come back at uneven weights as the widget moves. Without the layer the
// scale is a transform on the geometry instead, so rasterisation happens
// once, at final resolution.
//
// 2. CurveRenderer rather than the default geometry renderer. It rasterises
// curves analytically with its own antialiasing instead of tessellating them
// into AA'd triangles, which is what keeps sub-pixel detail smooth rather
// than stair-stepped. Qt 6.6+; this ships against 6.11.
Item {
id: root
property string pathData: ""
property real iconSize: Style.font.icon
property color color: Color.foreground
// Per-mark optical correction; see LOGO_SCALE in Modes.js for why equal
// nominal size is not equal apparent size.
property real opticalScale: 1.0
// Rounded so the mark is laid out on whole pixels. Thin strokes straddling a
// pixel boundary lose half their coverage to each side and read as grey.
readonly property int box: Math.round(iconSize)
implicitWidth: box
implicitHeight: box
visible: pathData !== ""
Shape {
id: shape
width: 24
height: 24
anchors.centerIn: parent
antialiasing: true
preferredRendererType: Shape.CurveRenderer
scale: root.box * root.opticalScale / 24
ShapePath {
// No fillRule set on purpose: these marks are authored for nonzero
// winding, which is the default, and LM Studio's bars are knocked out
// of its container by winding direction alone.
fillColor: root.color
strokeWidth: 0
PathSvg { path: root.pathData }
}
}
}
omarchy/smoido.claude-mode/Modes.js
+160
View File
@@ -0,0 +1,160 @@
.pragma library
// The four modes claude-mode understands, in the order the CLI menu lists
// them. Kept here rather than in either QML file because the bar widget needs
// the glyph and the panel needs the prose, and a second copy of this table is
// exactly the sort of thing that drifts.
// Material Design icons from the Nerd Font patch set, written as codepoints
// rather than literals so they survive any editor or transport that is not
// UTF-8 clean. All verified present in JetBrainsMono Nerd Font, the shell's
// default family.
//
// Deliberately none of them a robot: omarchy.agents already paints one, and it
// commonly sits in the same stretch of bar, where two identical robots read as
// one widget drawn twice. Same reason lmstudio is a server rack rather than a
// screen - omarchy.monitor owns that silhouette.
var GLYPH = {
anthropic: String.fromCodePoint(0xF0674), // sparkles - the real thing
openrouter: String.fromCodePoint(0xF0469), // router - remote gateway
zai: String.fromCodePoint(0xF015F), // cloud - hosted GLM plan
lmstudio: String.fromCodePoint(0xF048B), // server - local LM Studio
unknown: String.fromCodePoint(0xF0625) // question - nothing readable
}
// ---------------------------------------------------------------------------
// Brand marks
//
// The real logo for each provider, as SVG path data in a 24x24 viewBox so the
// whole set scales from a single number. Drawn with QtQuick.Shapes rather than
// set as font glyphs: three of the four have no Nerd Font pictograph at all,
// and a stand-in pictograph is both less recognisable and impossible to keep
// visually consistent with the marks beside it.
//
// Sources: Claude, OpenRouter and LM Studio from simple-icons; Z.AI from
// lobe-icons. All four are monochrome single-path marks, which is what lets
// them take the bar's foreground colour and follow the theme. Trademarks
// belong to their respective owners.
//
// LM Studio's mark keeps its rounded-square container: the four bars alone
// read as a hamburger menu, and the container is what makes it that logo. It
// relies on nonzero winding to knock the bars out, which is QtQuick.Shapes'
// default, so no fill rule is set anywhere here.
var LOGO = {
anthropic: "m4.7144 15.9555 4.7174-2.6471.079-.2307-.079-.1275h-.2307l-.7893-.0486-2.6956-.0729-2.3375-.0971-2.2646-.1214-.5707-.1215-.5343-.7042.0546-.3522.4797-.3218.686.0608 1.5179.1032 2.2767.1578 1.6514.0972 2.4468.255h.3886l.0546-.1579-.1336-.0971-.1032-.0972L6.973 9.8356l-2.55-1.6879-1.3356-.9714-.7225-.4918-.3643-.4614-.1578-1.0078.6557-.7225.8803.0607.2246.0607.8925.686 1.9064 1.4754 2.4893 1.8336.3643.3035.1457-.1032.0182-.0728-.164-.2733-1.3539-2.4467-1.445-2.4893-.6435-1.032-.17-.6194c-.0607-.255-.1032-.4674-.1032-.7285L6.287.1335 6.6997 0l.9957.1336.419.3642.6192 1.4147 1.0018 2.2282 1.5543 3.0296.4553.8985.2429.8318.091.255h.1579v-.1457l.1275-1.706.2368-2.0947.2307-2.6957.0789-.7589.3764-.9107.7468-.4918.5828.2793.4797.686-.0668.4433-.2853 1.8517-.5586 2.9021-.3643 1.9429h.2125l.2429-.2429.9835-1.3053 1.6514-2.0643.7286-.8196.85-.9046.5464-.4311h1.0321l.759 1.1293-.34 1.1657-1.0625 1.3478-.8804 1.1414-1.2628 1.7-.7893 1.36.0729.1093.1882-.0183 2.8535-.607 1.5421-.2794 1.8396-.3157.8318.3886.091.3946-.3278.8075-1.967.4857-2.3072.4614-3.4364.8136-.0425.0304.0486.0607 1.5482.1457.6618.0364h1.621l3.0175.2247.7892.522.4736.6376-.079.4857-1.2142.6193-1.6393-.3886-3.825-.9107-1.3113-.3279h-.1822v.1093l1.0929 1.0686 2.0035 1.8092 2.5075 2.3314.1275.5768-.3218.4554-.34-.0486-2.2039-1.6575-.85-.7468-1.9246-1.621h-.1275v.17l.4432.6496 2.3436 3.5214.1214 1.0807-.17.3521-.6071.2125-.6679-.1214-1.3721-1.9246L14.38 17.959l-1.1414-1.9428-.1397.079-.674 7.2552-.3156.3703-.7286.2793-.6071-.4614-.3218-.7468.3218-1.4753.3886-1.9246.3157-1.53.2853-1.9004.17-.6314-.0121-.0425-.1397.0182-1.4328 1.9672-2.1796 2.9446-1.7243 1.8456-.4128.164-.7164-.3704.0667-.6618.4008-.5889 2.386-3.0357 1.4389-1.882.929-1.0868-.0062-.1579h-.0546l-6.3385 4.1164-1.1293.1457-.4857-.4554.0608-.7467.2307-.2429 1.9064-1.3114Z",
openrouter: "M16.778 1.844v1.919q-.569-.026-1.138-.032-.708-.008-1.415.037c-1.93.126-4.023.728-6.149 2.237-2.911 2.066-2.731 1.95-4.14 2.75-.396.223-1.342.574-2.185.798-.841.225-1.753.333-1.751.333v4.229s.768.108 1.61.333c.842.224 1.789.575 2.185.799 1.41.798 1.228.683 4.14 2.75 2.126 1.509 4.22 2.11 6.148 2.236.88.058 1.716.041 2.555.005v1.918l7.222-4.168-7.222-4.17v2.176c-.86.038-1.611.065-2.278.021-1.364-.09-2.417-.357-3.979-1.465-2.244-1.593-2.866-2.027-3.68-2.508.889-.518 1.449-.906 3.822-2.59 1.56-1.109 2.614-1.377 3.978-1.466.667-.044 1.418-.017 2.278.02v2.176L24 6.014Z",
zai: "M12.105 2L9.927 4.953H.653L2.83 2h9.276zM23.254 19.048L21.078 22h-9.242l2.174-2.952h9.244zM24 2L9.264 22H0L14.736 2H24z",
lmstudio: "M14.025 0c3.492 0 5.237 0 6.571.68a6.24 6.24 0 0 1 2.725 2.724C24 4.738 24 6.484 24 9.975v4.05c0 3.492 0 5.237-.68 6.571a6.24 6.24 0 0 1-2.724 2.725c-1.334.679-3.08.679-6.571.679h-4.05c-3.492 0-5.237 0-6.571-.68A6.24 6.24 0 0 1 .68 20.597C0 19.262 0 17.516 0 14.025v-4.05c0-3.492 0-5.237.68-6.571A6.23 6.23 0 0 1 3.404.68C4.738 0 6.484 0 9.975 0zM7.688 16.313a1.313 1.313 0 0 0 0 2.625h11.625a1.313 1.313 0 0 0 0-2.625zm-3-3.75a1.313 1.313 0 0 0 0 2.624h11.625a1.313 1.313 0 0 0 0-2.624zm3-3.75a1.313 1.313 0 0 0 0 2.624h11.625a1.313 1.313 0 0 0 0-2.624zm-3-3.75a1.313 1.313 0 0 0 0 2.625h11.625a1.313 1.313 0 0 0 0-2.625z"
}
// Equal nominal size is not equal apparent size, so each mark is corrected by
// two measurements taken off a 200px render of it:
//
// bbox Z.AI and OpenRouter are wide-but-short marks - their ink spans only
// ~84% of the box height, so at equal nominal size they read smaller.
// ink LM Studio's filled container covers 69% of its box against ~38% for
// the other three, so at equal nominal size it reads much heavier.
//
// The correction is the geometric mean of both signals, which pulls the short
// marks up and the solid one down without letting either measure dominate.
var LOGO_SCALE = {
anthropic: 0.99,
openrouter: 1.12,
zai: 1.07,
lmstudio: 0.86
}
function logo(mode) {
return LOGO[mode] || ""
}
function logoScale(mode) {
return LOGO_SCALE[mode] || 1.0
}
var TITLE = {
anthropic: "Anthropic",
openrouter: "OpenRouter",
zai: "Z.AI",
lmstudio: "LM Studio"
}
var BLURB = {
anthropic: "Your subscription login. No gateway, no API key.",
openrouter: "Remote gateway, pay per token, any vendor.",
zai: "GLM coding plan on Z.AI's Anthropic endpoint.",
lmstudio: "Local LM Studio server. Offline and free."
}
var ORDER = ["anthropic", "openrouter", "zai", "lmstudio"]
// Everything except anthropic switches with a named preset.
function needsPreset(mode) {
return mode !== "" && mode !== "anthropic"
}
function glyph(mode) {
return GLYPH[mode] || GLYPH.unknown
}
function title(mode) {
return TITLE[mode] || "not installed"
}
function blurb(mode) {
return BLURB[mode] || ""
}
// Bar label. The preset matters more than the provider once you are on a
// gateway - "openrouter" tells you nothing you did not already know from the
// icon, where "cheap" is the thing you actually want to catch sight of.
function shortLabel(mode, preset) {
if (mode === "") return ""
if (mode === "anthropic") return "anthropic"
return preset !== "" ? preset : mode
}
function defaultRoot() {
return homeDir() + "/.claude-mode"
}
function homeDir() {
// Qt exposes no home path to a .pragma library, and the shell always runs as
// the owning user, so derive it the one way that holds in both places.
var url = Qt.resolvedUrl(".").toString()
var m = url.match(/^file:\/\/(\/home\/[^\/]+)\//)
if (m) return m[1]
m = url.match(/^file:\/\/(\/Users\/[^\/]+)\//)
return m ? m[1] : "/root"
}
// Presets from health.json that belong to one provider, default first so the
// list opens on the one `claude-mode <mode>` would have picked.
function presetsFor(health, mode, defaults) {
var all = (health && health.presets) ? health.presets : []
var out = []
for (var i = 0; i < all.length; i++) {
if (String(all[i].provider) === mode) out.push(all[i])
}
var preferred = defaults[mode]
out.sort(function (a, b) {
if (a.name === preferred) return -1
if (b.name === preferred) return 1
return a.name < b.name ? -1 : (a.name > b.name ? 1 : 0)
})
return out
}
// Mirrors default_preset_for() in the CLI.
var DEFAULT_PRESET = {
openrouter: "default",
zai: "zai",
lmstudio: "lmstudio"
}
function contextLabel(tokens) {
if (!tokens) return ""
if (tokens >= 1000000) return (Math.round(tokens / 100000) / 10) + "M context"
if (tokens >= 1000) return Math.round(tokens / 1024) + "k context"
return tokens + " context"
}
omarchy/smoido.claude-mode/Panel.qml
+1217
View File
File diff suppressed because it is too large Load Diff
omarchy/smoido.claude-mode/manifest.json
+18
View File
@@ -0,0 +1,18 @@
{
"schemaVersion": 1,
"id": "smoido.claude-mode",
"name": "Claude Mode",
"version": "1.8.0",
"author": "smoido",
"description": "Which provider Claude Code is pointed at, and a one-click switch between them",
"kinds": ["bar-widget"],
"entryPoints": {
"barWidget": "BarWidget.qml"
},
"barWidget": {
"displayName": "Claude Mode",
"description": "Active Claude Code provider and preset, with in-place switching",
"category": "Info",
"allowMultiple": false
}
}
presets/cheap.json
-18
View File
@@ -1,18 +0,0 @@
{
"provider": "openrouter",
"description": "Minimum cost. Everything on flash-class models with 1M context.",
"baseUrl": "https://openrouter.ai/api",
"auth": {
"mode": "vault",
"keyRef": "openrouter"
},
"models": {
"opus": "deepseek/deepseek-v4-pro",
"sonnet": "deepseek/deepseek-v4-flash-0731",
"haiku": "qwen/qwen3.7-flash",
"fable": "openai/gpt-5.6-luna-pro"
},
"subagentModel": "qwen/qwen3.7-flash",
"gatewayModelDiscovery": true,
"contextTokens": 1000000
}
presets/default.json
+6 -5
View File
@@ -1,18 +1,19 @@
{ {
"provider": "openrouter", "provider": "openrouter",
"description": "Daily driver. Cheap flash models on the hot tiers, GLM on fable.", "description": "Daily driver. Flash-class models on the hot tiers, GLM on fable.",
"baseUrl": "https://openrouter.ai/api", "baseUrl": "https://openrouter.ai/api",
"auth": { "auth": {
"mode": "vault", "mode": "vault",
"keyRef": "openrouter" "keyRef": "openrouter"
}, },
"models": { "models": {
"opus": "deepseek/deepseek-v4-flash-0731", "opus": "z-ai/glm-5.3-flash",
"sonnet": "qwen/qwen3.7-flash", "sonnet": "deepseek/deepseek-v4-flash-0731",
"haiku": "openrouter/free", "haiku": "openrouter/free",
"fable": "z-ai/glm-5.2" "fable": "z-ai/glm-5.3"
}, },
"subagentModel": "inherit", "subagentModel": "inherit",
"gatewayModelDiscovery": true, "gatewayModelDiscovery": true,
"contextTokens": 1000000 "contextTokens": 1000000,
"configured": false
} }
presets/lmstudio-qwen.json
-21
View File
@@ -1,21 +0,0 @@
{
"provider": "lmstudio",
"description": "Local Qwen3.6-35B-A3B uncensored (heretic, MTP-preserved). Clean chat template - verified working with tools + streaming.",
"baseUrl": "http://127.0.0.1:1234",
"auth": {
"mode": "literal",
"token": "lmstudio"
},
"models": {
"opus": "qwen3.6-35b-a3b-uncensored-heretic-native-mtp-preserved",
"sonnet": "qwen3.6-35b-a3b-uncensored-heretic-native-mtp-preserved",
"haiku": "qwen3.6-35b-a3b-uncensored-heretic-native-mtp-preserved",
"fable": "qwen3.6-35b-a3b-uncensored-heretic-native-mtp-preserved"
},
"subagentModel": "qwen3.6-35b-a3b-uncensored-heretic-native-mtp-preserved",
"gatewayModelDiscovery": true,
"extraEnv": {
"CLAUDE_CODE_ATTRIBUTION_HEADER": "0"
},
"contextTokens": 262144
}
presets/lmstudio.json
+8 -7
View File
@@ -1,21 +1,22 @@
{ {
"provider": "lmstudio", "provider": "lmstudio",
"description": "Local KAT-Coder v2.5. NOTE: its chat template asserts message order - if you hit 'Unable to generate parser for this template', use the lmstudio-qwen preset.", "description": "Local LM Studio server. Set up picks the model from whatever the server actually has installed.",
"baseUrl": "http://127.0.0.1:1234", "baseUrl": "http://127.0.0.1:1234",
"auth": { "auth": {
"mode": "literal", "mode": "literal",
"token": "lmstudio" "token": "lmstudio"
}, },
"models": { "models": {
"opus": "kwaipilot_kat-coder-v2.5-dev", "opus": "qwen3.6-35b-a3b-uncensored-heretic-native-mtp-preserved",
"sonnet": "kwaipilot_kat-coder-v2.5-dev", "sonnet": "qwen3.6-35b-a3b-uncensored-heretic-native-mtp-preserved",
"haiku": "kwaipilot_kat-coder-v2.5-dev", "haiku": "qwen3.6-35b-a3b-uncensored-heretic-native-mtp-preserved",
"fable": "kwaipilot_kat-coder-v2.5-dev" "fable": "qwen3.6-35b-a3b-uncensored-heretic-native-mtp-preserved"
}, },
"subagentModel": "kwaipilot_kat-coder-v2.5-dev", "subagentModel": "qwen3.6-35b-a3b-uncensored-heretic-native-mtp-preserved",
"gatewayModelDiscovery": true, "gatewayModelDiscovery": true,
"extraEnv": { "extraEnv": {
"CLAUDE_CODE_ATTRIBUTION_HEADER": "0" "CLAUDE_CODE_ATTRIBUTION_HEADER": "0"
}, },
"contextTokens": 262144 "contextTokens": 262144,
"configured": false
} }
presets/zai.json
+5 -4
View File
@@ -7,10 +7,10 @@
"keyRef": "zai" "keyRef": "zai"
}, },
"models": { "models": {
"opus": "glm-5.2", "opus": "glm-5.3",
"sonnet": "glm-5.2", "sonnet": "glm-5.3",
"haiku": "glm-4.7", "haiku": "glm-4.7",
"fable": "glm-5.2" "fable": "glm-5.3"
}, },
"subagentModel": "inherit", "subagentModel": "inherit",
"gatewayModelDiscovery": false, "gatewayModelDiscovery": false,
@@ -18,5 +18,6 @@
"API_TIMEOUT_MS": "3000000", "API_TIMEOUT_MS": "3000000",
"CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": "1" "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": "1"
}, },
"contextTokens": 1000000 "contextTokens": 1000000,
"configured": false
} }