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>
- 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
Replace the deepseek env redirect with a deepseek-patch-json recipe: a per-session DSH_HOME holding a generated cordis.patch.yml that declares a codeman-custom openai-completions provider over the installed profile. Ported from PR #752 commits 941df574 and 4bb1bfe9, minus the core context-limit changes.
- share one context-length resolution (discovered figure, endpoint default fallback) across the routes and boot recovery, with a recovery test
- record listed per-slot context figures with a slot count of 1 and force the slot-count fallback once per rediscovery
- open up the codex model launch token for qwen3-style ids (model-plain-colon), pinned by a spawn-golden case
- pass the full discovered window to opencode as limit.context
- translate the picker search strings and document the discovery precedence
- opencode: write limit.context/output from the discovered context length (#749)
- codex: override the native launch model when a custom endpoint is selected (#750), and set model_context_window
- model picker: add a search/filter box with a no-matches line
DeepSeek Harness's own bundled provider module
(@deepseek-ai/dsh-llm-deepseek) builds its request URL as
`${DEEPSEEK_BASE_URL}/chat/completions` with no `/v1` insertion of its
own (its real public API, https://api.deepseek.com, expects the
caller's base URL to already carry any needed prefix), while
llama-swap/llama.cpp only ever serves the OpenAI-conventional
`/v1/chat/completions`.
Confirmed two ways:
- Installed the real @deepseek-ai/dsh package (all its actual
published dependencies) into a scratch dir purely to read
dsh-llm-deepseek's source: `fetch(`${connection.baseURL}/chat/
completions`, ...)`, baseURL read straight from DEEPSEEK_BASE_URL —
the same grep-the-real-source bar pi/grok's fixes were held to.
- Live against the test-picker's llama-swap: `POST <baseUrl>/chat/
completions` -> 404, `POST <baseUrl>/v1/chat/completions` -> 200,
same endpoint. dsh's own error template ("DeepSeek API error (HTTP
${status})") reproduces the originally-reported
"dsh: HTTP_404: DeepSeek API error (HTTP 404)" exactly.
- New registry field `appendV1Suffix` (env kind only, deepseek's entry
alone — claude/gemini must NOT get it, since claude was already
confirmed working against the unmodified baseUrl). When set,
buildCustomModelInjection runs endpoint.baseUrl through the same
withV1Suffix() helper configDir-kind CLIs (pi/grok/codex) already
use, instead of writing it verbatim.
Not yet re-run end-to-end through a real dsh binary — no install
available in this environment (not in PATH, and the test-picker
container doesn't bundle it) — so this is source-confirmed and
live-verified at the HTTP level, not yet promoted to "verified"
alongside claude/opencode/pi/grok/omp. Docs (custom-model-endpoints.md,
the plan doc's confidence table, the wiki page, CLAUDE.md) all updated
to reflect this precisely rather than leaving the old "root cause not
identified" claim in place.
2 new/updated tests for the /v1 suffix (including idempotency against
a baseUrl that already ends in /v1) plus a corrected mock-server
contract test. Typecheck/lint clean; full suite shows no new
regressions.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RqZeHrRS6DYcGcGX2p9EwG
A fresh, isolated CLAUDE_CONFIG_DIR (used to keep an injected API key
away from a stored claude.ai OAuth login) looks like a brand-new Claude
Code profile to the CLI, so it replays its ENTIRE first-run sequence on
every single launch: the theme picker, the security-notes screen, the
per-project "trust this folder?" dialog, and (running with
--dangerously-skip-permissions) a one-time bypass-permissions warning —
confirmed live, none of which a real, already-onboarded profile shows
again.
- New registry-declared env-kind field `skipFirstRunPrompts` (alongside
apiKeyTrustFile, which it reuses) — claude's entry only, carried
through buildCustomModelInjection (pure) into
applyCustomModelInjection (IO).
- seedFirstRunOnboardingState(): merges hasCompletedOnboarding: true and
this session's own projects[workingDir].hasTrustDialogAccepted: true
into the same <configDir>/.claude.json the API-key trust file already
writes to — other projects and other fields on this session's own
entry are left untouched.
- seedSkipBypassPermissionsPrompt(): merges
skipDangerousModePermissionPrompt: true into <configDir>/settings.json,
a separate file, same corrupt-tolerant merge behavior.
- applyCustomModelInjection() gains an optional workingDir parameter,
threaded from session.workingDir (dedicated apply route) /
resolvedCasePath (quick-start route) — boot recovery omits it
(a dialog already answered once needs no re-seed on the same,
persisted isolated directory).
Tests added at the pure-builder, IO-wrapper (including merge-preserves-
other-fields and corrupt-file-tolerance cases), and existing directory-
listing assertions updated for the new settings.json file. Typecheck/
lint/format clean; full suite shows no new regressions (baseline
pre-existing Windows-environment failures unchanged, 8 more passing
tests than before — the ones added here).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RqZeHrRS6DYcGcGX2p9EwG
The CLAUDE_CONFIG_DIR isolation from the previous commit fixed the cosmetic
auth warning but introduced a real regression: an otherwise-empty config
directory has none of a real profile's prior custom-API-key approvals, so
Claude Code stops at an interactive 'Detected a custom API key - use it?'
prompt on every single launch. Confirmed live. With nobody at a TTY to
answer, the prompt's own default ('No') silently refuses the very key this
feature just injected, which looks like the endpoint being ignored.
Adds apiKeyTrustFile to the env-kind customModelInjection capability shape
({relPath, shape: 'claude-api-key-responses'}), set on claude's entry to
{relPath: '.claude.json', shape: 'claude-api-key-responses'}. The apply step
merges customApiKeyResponses.approved: [apiKey] into
<isolatedConfigDir>/.claude.json - the exact field a real answered prompt
itself writes to (confirmed against a real ~/.claude.json after answering by
hand once), so this answers the prompt in advance rather than bypassing it.
Merges onto whatever the CLI already wrote into that file on an earlier
launch in the same isolated directory rather than overwriting it; a missing
or corrupt file is treated as empty rather than failing the apply.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RqZeHrRS6DYcGcGX2p9EwG
Addresses two live-validation findings on the Run-menu custom-model picker:
1. Both claude.ai and ANTHROPIC_API_KEY set warning. Claude Code still
coexists an OAuth login with an injected ANTHROPIC_API_KEY in the same
config directory and warns about it (confirmed cosmetic - the API key
wins for actual requests, verified via a real session's own API Usage
Billing line). A custom-model claude session now gets an isolated
CLAUDE_CONFIG_DIR (registry-declared via a new configDirVar field, empty,
no files written into it) so there is nothing to conflict with. projects
is symlinked (junction on Windows) back into the real config dir so the
response viewer, subagent windows and Read My Mind keep working for that
session, best-effort.
2. Context-window overflow. Claude Code assumes a large default context
window for a model id it doesn't recognise and never compacts, so a
custom endpoint's real, much smaller context (verified live: a 400
exceeding a 16384-token llama-swap model with a stock ~33.7K-token system
prompt) silently overflows. Discovery now also learns each model's real
context length from llama.cpp/llama-swap's GET /props?model=<id> (n_ctx),
but ONLY for a model llama-swap's own /v1/models response already marks
status.value === 'loaded' - never an unloaded one, since llama-swap
treats ?model= as a routing hint and probing an unloaded model risks
triggering an actual, slow, GPU-swapping load as a side effect of
read-only discovery. A server with no status field at all gets no
enrichment rather than a guess; a model not probed this round keeps its
previously-learned value until it disappears from the list entirely.
Stored per model (CustomModelHost.modelContextLengths) and applied via a
new contextLengthVar registry field, set to
CLAUDE_CODE_MAX_CONTEXT_TOKENS for claude.
Both new fields live on the existing env-kind customModelInjection
capability shape, declared only on claude's registry entry - every other
CLI's injection is unaffected (pinned by test).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RqZeHrRS6DYcGcGX2p9EwG
Custom Model Endpoint Profiles (#393) let a session point its CLI at a
custom OpenAI-compatible endpoint by injecting env vars or a config file
and restarting the CLI in place. Review of the apply path found four
things, two of them destructive. This lands all four plus the smaller
items from the same review.
1. Clearing a selection did not clear it. The injected vars reach the CLI
via `tmux setenv`, which persists at the tmux-session level and is
inherited by `respawn-pane` (measured: `setenv FOO bar` survived two
successive `respawn-pane -k`), so deleting the keys from the session's
envOverrides relaunched the CLI still pointed at the old endpoint, and
for the configDir kinds at a HOME/CODEX_HOME/GROK_HOME that had just
been deleted. `Session.setCustomModel()` now reports the removed keys,
queues them (`_pendingEnvUnsets`), and `RespawnPaneOptions.unsetEnvKeys`
carries them into `applyEnvOverrides()`, which `setenv -u`s them before
re-applying the live overrides, on the same path that already unsets
the legacy CLAUDE_CODE_EFFORT_LEVEL. Verified on a private tmux socket
that `setenv -u HOME` hands the next respawn the global HOME back.
2. Applying a model to a local claude session killed the pane. The
relaunch was `claude --session-id <id>` and Claude refuses an id that
already has a transcript, and unlike the dead-pane respawn this one
kills a working pane first. `restartCli()` now pins the live
conversation id as the resume id for that respawn when the CLI's launch
declares a `fallback` chain, which renders the same
`--resume <id> || --session-id <id>` shape the docker and remote pane
commands use. Gated on the registry shape, not the CLI id: an entry
whose resume id is minted by the CLI itself never declares that chain.
3. pi, omp and grok wrote their config file and then launched without the
`--model` that selects it, so the file was ignored. The registry entry
now declares `customModelInjection.launchModel` (`custom/{modelId}` for
pi and omp, grok's `[model.codeman-custom]` block name), the builder
renders it, and `_withCustomModelLaunchModel()` applies it onto the
respawn options through `legacyConfigField`, leaving the stored
<Mode>Config untouched so a clear falls back to the user's own model.
A model id the CLI's `model` token pattern cannot carry is refused
with a 400 rather than silently dropped by the argv engine.
4. Remote (SSH) and Docker sessions reported `restarted: true` and changed
nothing: their `restartCli()` reattaches the durable tmux rather than
relaunching the agent, and the env lands on the local pane. Both are
refused with a 400 until those paths are plumbed.
Smaller items from the same review:
- The selection survives a Codeman restart as the disk-only `__customModel`
bookkeeping (endpoint, model, injected key NAMES, config dir, launch
model; never the values, which carry the API key). Recovery re-derives
the values from the endpoint store through the same apply path the route
uses and keeps the bookkeeping even when the endpoint is gone, so a
later clear still has keys to unset.
- Discovery goes through `webviewFetch()`, so the RESOLVED address is
judged by the same egress guard the web-tab proxy uses, and `baseUrl`
reuses `webviewUrlSchema` (http(s) only, no embedded credentials,
link-local and cloud-metadata addresses refused). undici's `fetch failed`
wrapper is unwrapped so the user sees the ECONNREFUSED underneath.
- `custom-model-hosts.json` is written 0600 via tmp+rename, the per-session
config dir 0700/0600 (pi and omp embed the key literally), and that dir
is removed with the session.
- `PR.md` is gone from the repo root and the design doc moved to
`docs/custom-model-endpoints-plan.md` with the LAN address and the
personal name scrubbed; every reference follows. The guide's `authStyle`
text matches the shipped schema (`bearer | api-key`, default `bearer`)
and says that `customModelEndpointsEnabled` is read by nothing until
the picker lands.
- `config/tsconfig.scripts.json` typechecks `scripts/test-local-llm-harnesses.ts`
(four real type errors fixed). It is not yet wired into `npm run typecheck`
because that line differs on master; adding `&& tsc -p config/tsconfig.scripts.json`
there is the one-line follow-up.
Tests: `test/session-custom-model-restart.test.ts` drives a real Session and
fails on the unfixed code for items 1 to 3; the route suite covers item 4
and the pattern refusal; `test/tmux-manager.test.ts` pins that the unsets
run before the overrides and that a shell-metachar key never reaches tmux.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Rewrites scripts/test-local-llm-harnesses.mjs -> .ts to read the live CLI
registry (enabledClis()) and call the real production
buildCustomModelInjection()/applyConfigDirInjection() instead of keeping a
second hand-maintained copy of every CLI's env/config shape. A future
registry change (new CLI, edited env var, fixed config template) is now
picked up automatically with zero edits to this script; only the one-shot
invocation flags (info the registry genuinely doesn't model) stay in a
small hand-maintained ONE_SHOT table, and a registry CLI with no entry
there reports UNKNOWN rather than being silently skipped.
Extracted src/custom-model-injection-apply.ts (applyConfigDirInjection/
removeConfigDir) so the production route and this script share one
implementation instead of two.
Full end-to-end run against a real llama-swap server, inside a
codeman/agent:llm-test Docker image with all 9 CLI binaries:
- claude, opencode, pi, grok, omp: PASS, real "hello world" replies
- codex: confirmed FAIL for a real protocol reason, not a bug — it only
speaks the Responses API since Feb 2026, which llama.cpp/llama-swap
don't implement
- gemini: confirmed FAIL, unresolved after real investigation — an
undocumented GATEWAY AuthType gemini-cli selects once
GOOGLE_GEMINI_BASE_URL is set rejects every auth-key format/override
tried
- deepseek: reaches the server (env vars are read) but gets a consistent
HTTP_404; root cause not identified, documented as best-effort/unknown
- antigravity: SKIP, no known mechanism (unchanged)
Two real bugs found and fixed along the way (grok, pi/omp registry
entries in stock.ts): grok's original recipe (env vars) was flat-out
wrong, not just unverified — the real mechanism is a config.toml
[model.<name>] block redirected via GROK_HOME. pi/omp's PI_CONFIG_DIR
does nothing for either (grepped pi's entire bundled source — the string
appears nowhere); the real redirect is the child process's own HOME, and
both need `models` as an array of {id} objects, not an object keyed by
id (silently loaded zero models otherwise).
deployment_plan.md, PR.md, docs/custom-model-endpoints.md, and CLAUDE.md
updated with the final confidence table reflecting all of the above.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017HqNWfmtBU2KN29SvSVWB3
Point any Codeman-supported harness (Claude, opencode, Codex, Gemini, Pi,
Grok, DeepSeek, OMP) at a custom OpenAI-compatible endpoint instead of its
native cloud backend, for a given session. Covers local hardware (llama.cpp,
Ollama, vLLM, DGX Spark, Strix Halo) and cloud (Azure AI Foundry, OpenRouter).
Off by default (customModelEndpointsEnabled, synced, default OFF).
- Registry: capabilities.customModelInjection per CLI entry (env /
configContentEnv / configDir / unsupported kinds)
- Pure injection builder (custom-model-injection.ts) turning an endpoint +
model id into the real env vars / config content per CLI
- Endpoint store + CRUD routes (custom-model-hosts.ts,
custom-model-routes.ts), discovery via GET /v1/models, SSRF-guarded
- Session integration: Session.setCustomModel()/restartCli()
(POST /api/sessions/:id/custom-model), reusing the existing
respawn-pane -k primitive to restart the CLI process with new env
- Multi-user hardening: every new redirect-capable env var added to its
CLI's privilegedEnvKeys, closing a pre-existing gap where several were
already reachable via the generic envOverrides field's prefix allowlist
- Standalone scripts/test-local-llm-harnesses.mjs: spawns real CLI binaries
against a real endpoint outside the web UI, independent of tmux/sessions
- Mock-server contract tests (test/fixtures/mock-openai-server.ts) replaying
every CLI's injected values through a real HTTP shape
Real end-to-end validation against a live llama-swap server (inside a
codeman/agent:llm-test Docker image with all 9 CLI binaries) found and
fixed three real bugs before they shipped:
- Codex's config.toml schema was wrong ([model].default table instead of
a top-level model string + [model_providers.custom]); fixing it then
surfaced a genuine, documented protocol incompatibility (Codex only
speaks the Responses API since Feb 2026, which llama.cpp/llama-swap
don't implement)
- Claude Code's async session-title-generation call validates
ANTHROPIC_DEFAULT_HAIKU_MODEL against its own internal model list and
hangs the whole -p invocation on an unrecognized name; documented for
chunk 6, worked around in the standalone script only (--bare is NOT
safe for a real interactive session, which needs hooks)
- The discovery route's authStyle: 'both' option (send both Authorization
and api-key headers) reliably hung a real server; removed the option
entirely rather than just changing the default
Status: draft. Chunk 6 (frontend toolbar/settings UI) not yet built — see
PR.md and deployment_plan.md for the full chunk breakdown and confidence
table.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017HqNWfmtBU2KN29SvSVWB3
Codeman maintainer