Found with knip (exports, types, class members) and confirmed by a text
search of src, test, scripts and packages before each removal.
- Per-CLI get<Cli>NotFoundMessage() (nine of them) with their message
constants: missingCliMessage() builds that text from the registry now.
- Registry leftovers: renderCliCommand, asCliId, cliIds, isCliRunnable,
createCliResolverForTest, profileExists, and two name lists that only
repeated schema enums.
- A task-description cache in Session that was written on every chunk of
terminal output (one regex pass per line) and never read since the
subagent watcher takes descriptions from transcripts. The cache module,
its pattern and both parse methods go with it.
- Session.parentAgentId / childAgentIds: never set by anything, so the two
optional SessionState fields were never on the wire.
- Unreferenced methods on TmuxManager (pipe-pane pair, syncMouseMode),
RespawnController, StateStore.recoverFromBackup (load() already falls
back to the .bak file), TeamWatcher, SseStreamManager and the AI
checkers' getErrorVerdict.
- updateCaseEnvVars: the settings.local.json env path that envOverrides
replaced.
- Re-exports in the utils, cli-registry, routes and prompts barrels that
no importer used, and unused test helpers (MockStateStore, the mobile
throttle helpers and others).
Eight mobile tests imported the WebServer type from a path that does not
exist (../src instead of ../../src); type-only, so nothing failed.
Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
The two commits of #803 that came after 1.51.0 was cut. The earlier commits (every model that fits, Muse Glimmer 30B and Ornith 1.0 9B) shipped in 1.51.0. No changeset: the next release names it.
Three corrections to the row added in 5396e33b, none of which changes where it is offered:
- Its peak. The run gave the process footprint's LIFETIME maximum, read at the end: 3.0 to
3.3 GB above what was held then. That is a floor, because a larger transient during the
earlier, longer prompts would not show in it. The row now takes the first measured row's
5 GB: 36.1 GB at 65,536 tokens (0.960 of a 48 GB Mac's estimate) and 39.0 GB at 131,072
(0.972 of what that Mac's runtime measures).
- What was compared. The two builds were not given the same prompt in OpenCode: the GGUF
build got a one-line task through the skill's sendmark (which adds its answer-file
sentence), the MLX build a plain one-line prompt. And the speeds to set side by side are
the two from the same 300-token request, 15.2 and 30.2.
- Runtime issue 18269: whether the memory fix its maintainer promised predates the pinned
runtime is not known, and the header says so.
Also a stale comment: the catalogue holds eleven models.
Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
Run on a MacBook Pro (M5 Pro, 48 GB, runtime 0.40.15) on 2026-11-12, after the GGUF build
this branch lists turned out to be unusable in agent use there: an OpenCode turn of one
line showed no answer after 7 minutes, twice, although the runtime alone wrote 15 tokens
per second.
The MLX build of the same model (`muse-glimmer:30b-mlx`, 19.1 GB) on that Mac:
- 32.6 tokens per second for a 1,500-token answer. The tag carries a draft model, and the
runtime logs speculative decoding with 0.34 to 0.62 of the drafts accepted.
- A correct tool call in under 3 seconds, and the same OpenCode turn finished in 45 seconds.
- Memory, read from the runner's own log line after every request: 37,900 to 45,100 bytes
held per token (the GGUF build computes to 7,072), prefix snapshots that filled their
8 GiB during twelve separate prompts, and a process peak 3.0 to 3.3 GB above what stayed
held.
So it gets a row of its own, the second with measured figures. They put it at 34.6 GB for
65,536 tokens: offered from 48 GB up, never on the 32 and 36 GB Macs, which keep the GGUF
row untouched (it is still the pick of a 32 GB Mac with a Max chip: whether it should be
is the maintainer's decision). It is never the pick of a Mac whose disk has room for
Qwen3.6 35B-A3B, and it has no `tested` entry: one turn is not a test.
Also written down from the same session: the session of many prompts the Qwen3.6 35B-A3B
row owed (5.6 GiB of prefix snapshots after twelve prompts, still growing; its figures are
unchanged), and what the two open runtime issues on the MLX tag say against what ran.
Not done: the row was not installed through the window, and the measurement used a second
instance of the pinned runtime, not Codeman's.
Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
Conflict in skills/codeman/SKILL.md (and its plugin mirror): the list of
modes with neither stop nor blocked gained jcode here and the omp
exception from #798 on the release branch; both are kept.
Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
The four edits promised in the review of #798:
- An omp alert could clear while omp's dialog was still up. omp items got
the same pane `capture` as Claude's, which the inbox re-reads with
Claude's numbered-dialog parser. omp's dialog has no numbered rows, but an
agent's plan above it does, so the list was parsed as the item's options,
and once it scrolled out of the frame the item was resolved as answered in
the terminal. An approval-only item now takes the frame once as `context`
and gets no `capture`, like a screen-read item (web/screen-blocked.ts), so
the module's `elicitation_complete` is the only thing that clears it. A
route test puts a numbered list above an omp-shaped dialog and fails on
the old code at both steps (options parsed, alert cleared).
- The reboot-restore install is dropped: the restore plan skips every CLI
whose transcript is not `claude-jsonl` (omp's is `omp-jsonl`), so it never
ran, and it sat after the start, when omp has already scanned its hooks.
- The docs that say which CLIs raise `blocked` and open alert-only items now
name omp: api-reference, CLAUDE.md, architecture-invariants, an entry for
`approvalHookFile` in cli-registry, and the skill with its plugin mirror
(verbs.md listed omp among the modes where `blocked` is a 400, one line
above the paragraph that says it works).
- Two comments: the schema's file-name rule is a pattern, not a closed list,
and the marker check takes the marker anywhere in the file.
Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
A second scripted run, on dsh 0.2.2-rc.1 with a freshly installed
dsh-TUI 0.19.0 under a scratch home: the model name reaches the endpoint
unchanged, the first request is 6,983 tokens with 27 tools, and a tool
round trip completes.
Two things that version does differently are written down where a reader
of the harness table looks: a setup guide opens in every session until a
person has finished it once (and its cursor is the readiness mark), and
its transcript file was renamed, so `last-response` is empty for it.
Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
An installed model knew one coding agent: the one it was installed with.
The launch check never asked about the agent, only about the endpoint
and the model, so nothing stopped a second one. What was missing is the
list, and a way to get an agent that is not on the Mac yet.
The install view gains `harnesses`: every catalogue harness the model in
use can be run with, each with how it stands (`ready`, `installable`,
`installing`, `setup-needed`), the literal install command, a readiness
mark for scripts, and the actions the server accepts for it. `cli` stays
the default agent, and names one that can be started whenever one can,
so a client that reads only `cli` never starts an agent that is gone.
`POST /api/local-model/agent { cli }` installs one more agent beside an
installed model. It is not a job and not an event of the reducer: the
state stays `installed` whatever the installer does, so a failed npm
install never blocks a session on a model that works. Same install slot
and same compiled stock command as the `agent` step, one at a time, the
outcome in memory only.
The catalogue:
- `pi` is no longer gated. Its two open questions were answered by a run
against a real model (the `name:tag` id reaches the runtime unchanged;
its first request is 2,096 tokens), and Codeman now passes it the
window. Its minimum is 65536 like the others: nothing below was run,
and no model row offers less, so no Mac's pick changes.
- `deepseek` is new. It is a launcher: its install command leaves it
with nothing to launch, so it is never reported installable (read off
`discovery.launcherProfile`, not the id) and a Mac is offered it only
where it was already set up. Its tool round trip was seen on a
scripted endpoint only; the header says how to gate it again.
- a row may carry `readyMark`; `opencode` has the measured one.
A switch no longer refuses for the agent a model was checked with when
another agent that can run it is on the Mac.
Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
A 16 GB Mac got no model: the two 9B builds of the catalogue need 10.2 GB at
65,536 tokens, 0.987 of the estimated budget, just past the 0.98 a tight fit
may reach.
New row `ornith:9b-q4_k_m` (Ornith 1.0 9B, MIT), a fine-tune of Qwen3.5-9B for
coding agents whose GGUF build has no vision projector: 5.63 GB against 6.55 GB,
so 9.3 GB at 65,536 tokens, 0.898 of that budget. SWE-bench Verified 69.4,
SWE-bench Pro 42.9, Terminal-Bench 2.1 43.1 on Ornith's card. Its config.json
and its GGUF header are those of Qwen3.5-9B (no prediction layer), so it shares
the figures of the two other 9B rows. Never run on a Mac: `estimated`.
- What changes for a Mac: 16 GB goes from no model to `limited` with this row,
tight. That is the only pick that changes: the row is ranked under MiMo V2.6
9B, so every Mac that could already run a 9B keeps its pick and gains an
alternative. With the budget a 16 GB Mac mini measured (10.7 GB) the pick is
MiMo as before.
- Why under MiMo although its own card's numbers are higher: the two vendors'
tables put their common base 7 points apart, and runtime issue 16992 (open)
was filed on the 35B sibling, which has the same renderer and parser. The
35B stays set aside.
- The 16 GB rule itself is untouched: same budget estimate, same 0.98, same
eight context checkpoints.
- Ranks renumbered 1 to 10; the pins of test/local-model-recommend.test.ts and
the table of docs/local-models.md follow.
Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
The "Run a local model" chooser showed the pick and four more models. It now
lists every catalogue model that fits the Mac, and the install accepts every
listed row.
- MAX_ALTERNATIVES goes from 4 to 11, above what the catalogue holds. It stays
a bound: test/local-model-catalog.test.ts fails for a catalogue that outgrows
it, so nothing drops off the list unseen.
- New row `muse-glimmer:30b-q4_k_m` (Meta, Apache-2.0): SWE-bench Verified 76.0,
SWE-bench Pro 51.2, Terminal-Bench 2.1 51.7 on Meta's card. Its K/V cache is
the smallest of the 27B class, so it is the first model of that class a 32 GB
Mac holds: 21.2 GB at 65,536 tokens, 0.970 of the estimated budget. Figures
computed from its config.json and its GGUF header by the llamacpp rule (13
full-attention layers, 39 sliding-window layers of 2,048 tokens, no
prediction layer); the sliding-window cache rule was read again in llama.cpp
b12111 (src/llama-kv-cache-iswa.cpp). Never run on a Mac: `estimated`.
- What changes for a Mac: 32 GB is no longer `limited`. On an M1 Max or M2 Max
the pick is Muse Glimmer; on a base or Pro chip the dense model is slow and
the pick stays MiMo V2.6 9B, now with the small-pick note. Every other pick
is unchanged.
- The "set aside" list was read again at its sources on 2026-11-10 and gained
ripple-preview:65b (every build requires runtime 0.40.16, one above the pin),
Ornith 1.0 (runtime issue 16992), nemotron-3.5-super:120b, qwen3.5:4b,
qwen3-coder-next and the MLX and DFlash tags of Muse Glimmer.
- The header and the guide now say that runtime issue 16992 has one diagnosed
cause in the tool parser every Qwen-family row shares (a tool argument that
quotes a whole tool call is cut off; the fix is runtime PR 17021, open).
Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
Review follow-up. omp declared `hooks: 'always'`, which every consumer reads
as Claude's full hook set, so an omp session picked up behaviour it cannot
back. omp goes back to `hooks: 'none'`; `capabilities.approvalHookFile` is now
a separate, narrower capability with its own predicate,
`approvalHookAvailableForMode()` (local, non-adopted sessions only).
- An explicit `stop` wait on omp is a 400 again; `blocked` still works.
- omp inbox items are alert-only (`answerInTerminal`, no `approvalId` on the
push), and only `permission_prompt` / `elicitation_complete` are accepted for
an approval-only session.
- The module is installed from POST /api/quick-start too (new and existing
cases, not Docker or remote) and on reboot-restore, after the launch
pre-flight on both create paths. Claude's settings.local.json block and the
Docker CLAUDE.md no longer land in omp workspaces.
- The generated module passes Bun's `tls: { rejectUnauthorized: false }` for an
https CODEMAN_API_URL (checked against a `--https` server), clears the alert
only when the last pending toolCallId resolves, sends reports in order,
carries omp's policy `reason` as `message`, and imports HookAPI from
`@oh-my-pi/pi-coding-agent/extensibility/hooks`.
- Every path component and the target file are lstat-checked.
- docs/omp-integration.md, verbs.md and the stock.ts comments updated.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
An omp pane asking for tool approval raised no tab alert, no desktop
notification and no Approvals Inbox item. omp inherited hooks:'none' from
agentDefaults(), so hooksAvailableForMode() answered false and
hook-event-routes never opened an item; there was nothing to resolve either.
omp is one of the few non-claude CLIs with real approval events
(tool_approval_requested / tool_approval_resolved), so Codeman can install a
hook module into <workspace>/.claude/hooks/pre/ - the directory omp scans -
on session create. It forwards them as permission_prompt and
elicitation_complete, the names the server already routes on, so the existing
alert, notification and inbox pipeline works unchanged.
The file name comes from the new capabilities.approvalHookFile field rather
than a branch on the CLI id (cli-registry-no-id-branching). stop/idle_prompt
are deliberately NOT reported: omp's idle stays screen-read via workDetect,
which the agent skill's verbs.md now says explicitly.
The plain preview already serves a workspace dotenv file in full to the same user, so refusing the edit protected nothing. The edit gate now passes allowDotenv to the sensitive-path check, which skips only the two dotenv patterns; every other credential pattern, a dotenv file inside a blocked tree and Codeman's own data-dir login file stay blocked. Downloads and attachments are unchanged.
jcode (1jehuang/jcode) joins the CLI registry as the `jcode` run mode:
Run menu, phone overview, TUI, quick-start and POST /api/sessions
(`jcodeConfig: { model?, resumeSessionId? }`), local, SSH and Docker
(dedicated agent-image layer), plus codeman doctor and install.sh
through the generated catalogue. Measured on a live jcode 0.98.1 pane.
- Work detection reads jcode's braille status row and `<n>>` composer;
alt screen with mouse strip and PageUp/PageDown paging like opencode.
- Every jcode TUI is a client of one shared per-user daemon whose tools
run with the daemon's env, so no agent skill injection and no JCODE_*
envOverrides prefix.
- New `capabilities.sharedDaemon`: a session kill spares descendants
outside the pane's process group, so closing the tab that started the
daemon no longer cuts every other jcode session off (measured).
- last-response answers an honest empty for an external CLI with no
transcript reader that the pane segmenter cannot read (jcode alone).
- GET /api/jcode/status via the generic registry resolver.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
The previous commit counted 4 rollback copies of the recurrent state for every GGUF with a
prediction layer, the runtime's default `draft_num_predict`. A tag can set its own in its
params layer, and the runtime then passes that as `--spec-draft-n-max`, which is what
`need_n_rs_seq()` turns into rollback copies. Read on 2026-11-08:
- qwen3.6:35b-a3b-mtp-q4_k_m: draft_num_predict 2, so 2 copies (was counted as 4).
- qwen3.8:27b-mtp-q4_k_m: draft_num_predict 4.
- qwen3.8:27b-q8_0: none in its params, so the runtime's 4.
The 35B-A3B GGUF row's overhead drops by 132 MB: 0.971 instead of 0.976 on a 36 GB M3 Pro or
M3 Max, the same pick. test/local-model-catalog.test.ts now carries the drafted tokens per row
instead of a yes/no prediction layer.
Also states what was verified about llama.cpp's checkpoint default: 8 from b6800 to b8200,
32 from b8250 together with a checkpoint every 8,192 prompt tokens (common/common.h at those
tags), in place of an inferred reason. Checked in the b12111 source as well: the draft context
of a Qwen-family MTP model caches the prediction block alone (`mtp_on_hybrid_qwen`, layers at
or above n_layer()), and the main context leaves it out, so the row counts that layer once.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The llama.cpp memory rule of the catalogue left out two allocations that the llama.cpp the
runtime bundles (b12111, the runtime's LLAMA_CPP_VERSION) makes for every GGUF row:
- Context checkpoints. For a model whose state cannot be rolled back (the recurrent state of
the Qwen-family rows, the sliding window of North Mini Code), llama-server keeps host copies
of that state, two or three per request, up to 32 per slot by default
(tools/server/server-context.cpp create_checkpoint). For the 27B that is 32 x 157 MB = 5 GB,
none of which was counted. The runtime now sets LLAMA_ARG_CTX_CHECKPOINTS=8 (read by
llama-server from its environment like LLAMA_ARG_CACHE_RAM; 8 was llama.cpp's own default
before it went to 32), and every llamacpp row counts 8 copies of its state.
- Multi-token prediction. The two -mtp tags and qwen3.8:27b-q8_0 carry a prediction layer
(nextn_predict_layers 1, read from the first bytes of each GGUF; the 9B tags and North Mini
Code have none). The runtime then always passes --spec-type draft-mtp --spec-draft-n-max 4,
so llama.cpp keeps 4 rollback copies of the recurrent state (need_n_rs_seq(),
llama-memory-recurrent.cpp) and the prediction layer's cache in the draft context's cache
type, f16 by default. The rows counted one state and that layer at q8_0.
What moves (test/local-model-recommend.test.ts, "every memory size", re-pinned):
- 16 GB: limited -> none. MiMo 9B at 65,536 tokens needs 10.2 GB, 0.987 of the 10.4 GB
estimate. With the budget the Mac mini measured (10.7 GB after the reserve) it is offered,
0.954, tight; that test is kept. With 4 checkpoints instead of 8 it would be 0.966 on the
estimate: this size sits on the edge, and only a run on such a Mac settles it.
- 18 GB: the same pick, 0.964 instead of 0.928.
- 36 GB: the same picks, 0.966 (27B on M4/M5 Max) and 0.976 (35B-A3B GGUF on M3).
- 48 GB on a Max chip: the 27B at 262,144 tokens is tight (0.853) instead of good.
- 32 GB with a measured budget of 0.74: North Mini Code at 131,072 tokens (was 262,144), the
27B at 65,536 on a Max chip (was 131,072).
- Every MLX row, and everything from 64 GB up, is unchanged.
test/local-model-catalog.test.ts recomputes the new rule from the architecture table and holds
the catalogue to LLAMA_ARG_CTX_CHECKPOINTS. The catalogue header, runtime.ts, the types,
docs/local-models.md, CLAUDE.md and architecture-invariants say the same. Read in the source,
not measured.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Q8 27B row counts a prediction layer its GGUF may or may not carry
(not checked, careful side), and the macOS notice on the runtime's page is
about Laguna XS 2.1; Laguna S 2.1 is set aside for its open runtime issues.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The 48 GB row now says the test was a short task and that MLX may keep up
to 8 GiB more in a long session, which brings the pair to the edge of the
budget. The check after the download says what it proves for each runner,
from the runtime's source: for a GGUF build both of its figures are the
buffers llama-server reports allocating, so the check covers the whole
load whatever the cache type; for an MLX build only the weights. And the
docs say where a new row's architecture goes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The catalogue had seven Qwen-family rows that all carried the memory
figures measured on one model (Qwen3.6 35B-A3B on the MLX runner), so a
9B or a dense 27B was judged with figures that do not describe it, two
rows had the wrong Mac download size, and every Mac below 32 GB got
"none".
Now eight rows, each with the runner a Mac really uses for its tag (a new
`runner` field: `mlx` or `llamacpp`) and memory figures of its own:
- the 35B-A3B MLX row keeps its measured figures and the tested pair;
- every llama.cpp row is computed from its model's config.json: a q8_0
K/V cache for the whole context (what the runtime now sets), its
recurrent or sliding-window state, 1.0 GB of runner and the 1 GiB host
prompt cache;
- the one other MLX row (Ornith 1.5 122B) is computed the same way,
scaled by what the measured row showed, plus the runner's 8 GiB of
prefix snapshots.
Sizes are the registry's Mac pull totals (2026-11-08, lowercase ids
verified there). What each memory size gets: 8 GB none; 16 to 32 GB a 9B
(limited); 36 GB Qwen3.8 27B or the 35B-A3B as GGUF; 48 GB the tested
pair (or the 27B on a Max chip); 64 and 96 GB the 27B or the 35B-A3B;
128 to 512 GB Ornith 1.5 122B. Nothing but the one row was measured.
The catalogue test recomputes every computed figure from a table of
architectures and lets two rows share figures only when they share the
architecture and the runner. The rule tests of recommend() now run on a
frozen copy of the 2026-11-06 catalogue, so a weekly catalogue update
changes only the per-memory table and the real-Mac tests.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This reverts commit 0aa34da8, which made the codeman agent skill default on here as well. That change lives in feat/agent-skill-default-on, together with the self-ignoring copy and the reference-folder symlink check it needs.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
agentSkillEnabled now reads an absent key as ON, the way
workspaceHooksEnabled does: only an explicit false turns it off. A new
session therefore finds the codeman skill in its folder without a
setting first, and with it the verbs that hand work to other CLIs and to
the local model on this Mac. Settings shows the switch on for the same
absent key.
The injection itself is unchanged: add-only, marker-owned, never through
a symlink, never for remote sessions or a plain shell. Turning the
setting off still removes nothing already written.
The quick-start test now pins the new default (injected on an absent
key, nothing after an explicit false), and the docs that said "default
off" say when that changed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Conflicts resolved as on test/desktop-macbook: config/test-suites.ts keeps both browser tests (desktop-connect-page and local-model-ui), and the CLAUDE.md state-files line keeps #778's cli-installed entry with the local model's local-model.json and local-model/ entries.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The catalogue rows carry the figures measured on a Mac (the MLX build of
qwen3.6:35b-a3b is 23,105,798,956 bytes, about 37 KB held per context token,
about 3 GB of runner overhead, about 5 GB more for a moment while a long
prompt is processed). A model's need is now its peak. Rows that were not
measured use the same per-token and overhead figures and say `estimated`.
qwen3.6:35b-a3b gets the tested entry for an Apple M5 Pro with 48 GB and
OpenCode, with its measured 67.7 tokens per second, so that row may say
"Tested on a Mac like this".
The ranking changes with it: a pair tested on a Mac like this one comes
first, then speed (ok, unknown, slow), then model rank, harness rank and the
larger context. The fit grade decides whether a row is listed and what it
says, it no longer sorts, so a model is offered at the largest context it
fits at. recommend() takes the budget the runtime measured
(`budgetSource: measured`), and modelFitsBudget() judges one model against it.
Candidate rows now always carry `needSource`, `harnesses`, `installBytes`,
`installGb` and, where one applies, `tag`. `kvBytesPerToken` and the prompt
cache tiers are gone.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Keep the 1.49.0 API meaning: liveWorktree and root=live stay inside the
session folder, and an outside worktree is published as outsideWorktree
and walked only with root=live-any.
Refuse a symlinked .git folder as the main worktree, sensitive files
inside a followed worktree, and any outside worktree while attachment
confinement is on or around a blocked tree. The file routes re-check
the guard settings on every request, and a throwing rule check refuses.
The panel keeps its Back to worktree button while pinned to the session
folder.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The core of the read-only recommendation: what this Mac is, what the catalogue
knows, and which model fits. Nothing is downloaded, installed or written.
- hardware.ts: one bounded async spawn of /usr/sbin/sysctl with seven keys,
without -n, parsed as "key: value" lines by the pure parseSysctl(). Free disk
of the data dir through statfs behind the bounded path probe. Fallbacks from
node:os, a 60 s cache, never throws, never system_profiler. A wired limit of
0 means "not set".
- local-model-catalog.ts: seven models, three harness rows (pi gated) and the
chips table as a compiled constant, validated by a Zod schema on first use.
No override file. Model ids are name:tag in lowercase without a slash; every
model carries paramsB and an (empty) tested list. Chip names match on the
end of the chip string, so 'M4' is not the chip of 'Apple M4 Pro'.
- recommend.ts: device budget (the wired limit when above 0, else 2/3 of
memory below 36 GB and 3/4 from 36 GB, less the reserve), need = weights +
KV cache + 1 GiB + prompt cache, fit good up to 0.85 and tight up to 0.98,
speed ok, unknown or slow (a chip in no chips row is unknown), and the
ranking. Verdicts ok, limited (best model under 12 billion parameters) and
none with its reason. "Recommended" only for a pick with a tested entry for
this Mac, which no row has today.
- types/local-model.ts: comments brought in line with the code (chip matching,
budgetShare rounding, a budget of 0 on an Intel Mac).
- cli-registry-no-id-branching.test.ts: the guard also catches a comparison of
a `cli` field with a CLI id (plus one stray blank line Prettier refused).
Tests: test/local-model-format.test.ts, test/local-model-hardware.test.ts
(the captured output of two real Macs), test/local-model-catalog.test.ts,
test/local-model-recommend.test.ts.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The contract three builders code against in parallel. Nothing is recommended
yet: the route answers 501 and the catalogue is empty.
- src/types/local-model.ts: HardwareFacts, the catalogue rows, HarnessFact,
Candidate, Recommendation, LocalModelStatus, and LOCAL_MODEL_LABELS, the one
list of sentences the server may put into the payload (the zh-CN dictionary
is keyed by them).
- src/local-model/format.ts: formatGb, the one place a byte count becomes a
size a person reads (decimal GB), and memClassGb.
- src/config/local-model-catalog.ts: the Zod schema (model ids are name:tag,
lowercase, no slash) and an empty stock catalogue. No override file.
- Skeletons with their final signatures: hardware.ts, recommend.ts,
harness-facts.ts, routes/local-model-routes.ts (injectable dependencies),
public/local-model-ui.js with its script tag.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The parser and classifier run synchronously on the event loop over text a
pane, a container or an ssh host wrote, and one process could stall the
server for every session: 4 s for a 128 KiB argv of `/pi/` tokens, about a
minute for 2 MiB of socket rows.
The `cli.js` launcher rule leaves the signature regex for a per-word scan,
socket rows dedupe through a Set, the process row regex takes the `s` flag
(a bare CR made `$` fail and `\s+` give its spaces back, quadratic in a row),
argv and pane command are cut at FOREIGN_MAX_ARGV_CHARS, and one matcher per
scan answers each distinct argv once, since every pane of a hostile host can
claim the same process tree.
Measured on a private tmux server with the same 131 KiB argv: the largest
event-loop gap went from 3991 ms to 45 ms. The worst scan the caps allow
takes about 80 ms. Socket rows stay uncapped on purpose (see the docs).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A remote host whose login shell is tcsh came up on 1.48.0, when the local
sh rewrote ${SHELL:-/bin/sh} to the Codeman host's shell. Since #770 the
launch reaches the far end verbatim and runs tcsh itself, and tcsh refuses
-i -l -c, -l -c and -i -l, so every stock launch left a dead pane.
The login flags are now picked on the far end, in the shape of
loginShellArgs(): set -- -i -l, cleared for csh and tcsh, then
exec "${SHELL:-/bin/sh}" "$@". The text is one brace group, so
cd <path> && { ...; } still starts nothing when the cd fails. bsd-csh takes
every flag combination, csh is listed because it is tcsh on macOS and the BSDs.
The version probe keeps its one statement: the login shell parses it, not
/bin/sh, so fish and tcsh refuse it either way, as they did on 1.48.0.
Measured with the real builder on a real sshd in a throwaway Debian container
(tcsh 6.24.07, bsd-csh, fish 3.6, bash) and through a fake sshd with
tcsh 6.24.10, fish, zsh, dash and bash, a real Codeman server on tcsh included.
The tcsh disclosure comes out of the docs and the changeset. They keep the two
limits 1.48.0 had too: tcsh has no export, so the PATH bootstrap only prints an
error, and the tmux probe needs a command executable, which Debian lacks.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The attach line now takes the rule #770 gave remote and docker launches: bashCArg() single-quotes it at both respawn-pane sites, so it reaches bash verbatim. quoteAdoptLaunch is retired, since its double quotes kept the same $ and backtick safety but lost a real newline. #510's no-cd condition stays.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Names tmux reads as target syntax (`@1`, `%1`, `=eq`, a bare `~`) passed the
has-session guard by name and then joined something else in new-session -t:
another session, or a fresh shell on the foreign server. They are refused at
discovery now, and take the existing skipped note.
One stopped tmux server emptied the whole local list and left a stuck client
per scan. Each server now gets a bounded ping with stdout at /dev/null first,
because a tmux client hands dups of its stdout to the server: a stopped server
holds our pipe open after the client is dead, so a watchdog around list-panes
cannot work. The probe runs in its own process group, all three transports
keep the rows read before a failure, and a missed ping becomes a note.
The attach and the probe pass -u, so a POSIX-locale container no longer garbles
the tab or lists a non-ASCII name as underscores. A docker attach gets the same
TERM and COLORTERM an owned container's exec gets.
A missing tmux no longer reads as "session is gone", and the launch line's `$`
now reaches bash unexpanded, so the ssh attach's PATH line is expanded at the
far end, as the probe's already was.
fix-plan/write, fix-plan/read and ralph-prompt/write refuse an adopted session:
its workingDir is the neutral /tmp, so they read and wrote a predictable name
in a folder anyone can write to.
FOREIGN_MAX_SOCKETS is gone: it capped nothing, and a real cap in glob order
would never reach `default` on a box with many `codeman-*` sockets (this one
has 212). Closing an adopted tab re-reads the outside-sessions list.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The setting told users the sandbox and approvals stay on. With it on, a
command in Codex's sandbox can call Codeman's API, which can start a session
outside the sandbox with no approval prompt. Measured on codex 0.162.0 with
`codex sandbox`: a touch the sandbox refused landed through POST /api/sessions,
and from inside a Codeman pane it landed with a password set too, because a
pane inherits CODEMAN_PASSWORD. Without that password the same call gets 401.
The settings row, the wiki, the skill and the changeset now say so, the wiki
Security page and security-architecture name it, and the docs scope the claims
to what the code does: app-wide, local launches only, not cron jobs, and
--no-daemon is a local launch's flag. Agent-CLIs gains the --no-daemon
trade-off. A test pins that no surface sells the sandbox as still in charge.
Text only: the switch, the clamp and the launch are unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The owner's answer to the PR's open question: the switch does not also pass
--sandbox workspace-write. Settings, wiki, skill and release note now state
the consequence: a folder Codex runs read-only keeps the network closed and
asks per command. A golden test pins the decision by name, and the stale
launch-defaults comments now name networkAccess.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Brings the 1.33.2 based adoption feature up to master. Conflicts were
resolved by keeping master's intent and re-applying the feature on top:
both sides kept for imports, the script tag and the stylesheet blocks; the
trust-dialog guard keeps master's stopped check plus the adopted refusal;
the Ralph and image watchers keep master's remote branch plus the adopted
skip; the close confirmation keeps master's worktree option plus the
adopted wording.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Conflict in src/session-cli-registry-bridge.ts: kept the PR's gateVersion() (gates measured against the gated CLI's own version) and added master's Discord startup hint engine value.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The gate comment in terminal-ui.js said a stale-false flag falls through
to the PageUp/PageDown fallback, which is false for pi since the second
commit of this PR removed that fallback. The FOOTGUN comment above it made
the same claim. Say what happens instead.
"Every pi below 1.0.0 emits no mouse enables" was too absolute: regular
was the default there, but fullscreen exists from 0.84.0. Measured with
--tui-mode fullscreen under tmux on 0.84.1, 0.85.0, 0.86.0 and 0.99.2:
alternate_on=1, mouse_any_flag=1, mouse_sgr_flag=1. Default settings on
0.99.2 report all 0.
Document that Wheel Scrolls Local History leaves the wheel dead in
fullscreen pi (wiki, integration guide, changeset), add a pi block to
Troubleshooting, cross-reference the accepted exception from the dead
gesture rule in architecture-invariants, and mark the pi-integration-plan
passages this PR reversed as superseded.
Comments and docs only, no runtime change.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
transcriptPageKeys applied to every pi, and a pi in `tuiMode: regular` binds
PageUp/PageDown to its editor. It emits no mouse enable, so the flag stays
false, and on a fresh pane the buffer is hollow: the wheel sent PageUp, the
caret of a multi-line draft moved, and the next keystroke landed on the wrong
line. Measured on real pi 1.1.0, 1.4.0 and 1.5.0. Regular mode is not exotic,
it only takes one line in pi's settings. Telling the two modes apart would
need a new capability, since the paging gate takes no id, so pi has no
fallback: a fullscreen pi with "Wheel scrolls local history" on stays where
master left it.
altScreen, the half of the fix that does the work, was pinned by nothing:
reverting it left 1328 tests green. The strip decision table reads its
expected value from the same registry entry. pi now has its own block using
the bytes it really emits, which also pins that focus reporting survives.
The fullscreen default dates from pi 1.0.0, not 1.4.0, so the comments say
that, and the sites still calling pi narrow-strip or main-screen by default
are corrected. Where a comment listed CLIs it now names the capability.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
pi 1.4.0 is an alt-screen TUI that enables mouse tracking once at start.
Codeman's replay (tab switch, reload) drops that mode in xterm
(mouseTrackingMode none, baseY 0), so the wheel hit an empty local buffer
and stayed dead until a resize made pi redraw.
Same shape as copilot: altScreen strip-mux-and-mouse (the server records
the stripped DECSETs as cliMouseTracking, which survives the replay),
wheelForward always, transcriptPageKeys (PageUp/PageDown fallback for a
stale flag, measured on pi 1.4.0). _shouldForwardWheelToApp treats pi like
copilot. Registry pins and docs updated.
Tests: 2 new in terminal-scroll-routing (red before). Full run: 12989
passed, 5 red from the de_DE locale / XDG env (custom-model-run-menu-ui,
git-worktrees, edit-history-sync), all green under LANG=C.UTF-8.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
(cherry picked from commit 53e4136d9f4401e9250ac6a122599ffcb2a57e8a)
Consolidates the four contributor changesets, and rewords the em-dashes this release
added in comments and docs. The five in test/settings-save-failure-ui.test.ts stay:
they assert existing toast text.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Say that settings.yaml is shared only until the session's first settings write
(dsh saves by rename, which replaces the link), warn when a session's own DSH_HOME
does not answer and the server's home is linked instead, test
guardSessionConfigDirHome, and document the linkFromRealHome exception and what a
custom-model dsh session does not carry over.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Second review round on PR #755: skip non-absolute DSH_HOME candidates so a relative home can never mkdir under the server cwd; probe a session-supplied home with probePathKind in both async routes before the synchronous apply; test the session env-override lookup (seventh arg) incl. the self-link fall-through and the relative-home skip; reject '.' and '..' as realHomeFallback and enforce linkFromRealHomeDirs subset in the schema; fix the stale apply JSDoc and the two doc lines lagging the verified dsh run; state the shared/isolated split plainly.
- linkFromRealHomeDirs marks directory entries: missing dirs are created (0700) instead of leaving a dangling link that failed every prompt with ENOENT; missing file targets are skipped
- _restoreCustomModel rebuilds the bookkeeping from the CURRENT recipe (configDir + env-key union) so a selection saved under the old env recipe can still be cleared after upgrade; drop the dead session-home recovery lookup
- deepseek's patch now references CODEMAN_CUSTOM_MODEL_API_KEY, so an in-session /model switch back to the official route cannot carry the custom key to api.deepseek.com (privilegedEnvKeys updated)
- schema: realHomeFallback must be one relative segment; linkFromRealHome requires it and must not collide with fileName
- docs: maintainer-verified real-dsh run replaces the outstanding-verification wording; shares-vs-isolates note; plan table row refreshed
Codeman maintainer