feat(config): add UI control for agent action timeout (#3100)#3112
Merged
senamakel merged 3 commits intoJun 1, 2026
Conversation
) Adds a user-facing "Action timeout" setting so users running large local models can extend the tool/action wall-clock limit without hand-editing config files. Previously the timeout was env-var-only (OPENHUMAN_TOOL_TIMEOUT_SECS), cached immutably in a OnceLock. - tool_timeout: make the effective timeout runtime-mutable (AtomicU64 + set_tool_timeout_secs) with a pure resolver; the operator env var still wins when set, otherwise the persisted config drives it. Read fresh per tool call so changes apply without a restart. - config: add [agent].agent_timeout_secs (default 120, range 1-3600) plus config_get_agent_settings / config_update_agent_settings RPC (controller registry only). Seed the runtime value from config at startup. - UI: "Action timeout" section in AgentAccessPanel (Settings -> Agent OS access) with env-override notice; i18n keys in en + all 13 locales. - about_app catalog + coverage matrix + .env.example docs updated. - Tests: tool_timeout resolver/env-precedence, config ops apply/reject, schemas registration, json_rpc_e2e roundtrip, and Vitest panel coverage (load/persist/range-reject/no-op/env-override).
Contributor
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (3)
🚧 Files skipped from review as they are similar to previous changes (2)
📝 WalkthroughWalkthroughIntroduce a runtime-mutable agent action timeout (1–3600s, default 120s) with env override precedence, persisted config, RPC get/update endpoints, frontend Settings control, i18n strings, startup seeding, docs, and tests. ChangesAgent Action Timeout Configuration
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 minutes Possibly related PRs
Suggested reviewers
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
coderabbitai
Bot
added
feature
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 4
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@app/src/components/settings/panels/AgentAccessPanel.tsx`:
- Around line 230-239: The save-response handler can overwrite a newer user edit
because it only checks request order (timeoutSeqRef) rather than whether the
input has changed since the save began; capture the input value when initiating
the save (e.g., const originalInput = timeoutInput or use a ref), call
openhumanUpdateAgentSettings with parsed, and on success only call
setTimeoutInput and setTimeoutSavedNote if timeoutSeqRef.current === seq AND the
current input still equals originalInput (or ensure the current input ref
matches originalInput); still update setSavedTimeoutSecs when the save
completes, but avoid snapping timeoutInput back if the user has typed a newer
value since the request started (referencing timeoutSeqRef, setTimeoutInput,
setSavedTimeoutSecs, setTimeoutSavedNote, openhumanUpdateAgentSettings).
- Around line 92-107: The current Promise.all([...]) call couples
openhumanGetAutonomySettings() and openhumanGetAgentSettings() so any rejection
from the agent timeout RPC prevents updating autonomy state; split these calls
and handle errors independently so autonomy still renders: call
openhumanGetAutonomySettings() and update setLevel, setWorkspaceOnly,
setRequireTaskPlanApproval, setTrustedRoots, setAutoApprove inside its own
try/catch (or use Promise.allSettled and apply results.result for autonomy even
if agent fails), then separately call openhumanGetAgentSettings() and update
setTimeoutInput, setSavedTimeoutSecs, setTimeoutEnvOverride, setTimeoutMin,
setTimeoutMax inside its own try/catch and log or ignore agent errors without
returning early; keep the cancelled check semantics around both updates.
In `@src/openhuman/config/ops.rs`:
- Around line 1023-1031: get_agent_settings currently reads persisted config but
calls tool_execution_timeout_secs() which may still return an env/default value
until set_tool_timeout_secs() has been called elsewhere, causing
effective_timeout_secs to be stale; fix by computing the effective timeout from
the loaded config (or by calling
set_tool_timeout_secs(config.agent.agent_timeout_secs) prior to reading
tool_execution_timeout_secs()) so effective_timeout_secs reflects
config.agent.agent_timeout_secs; update get_agent_settings to use the config
value (or call set_tool_timeout_secs with config.agent.agent_timeout_secs) and
reference get_agent_settings, load_config_with_timeout,
tool_execution_timeout_secs, set_tool_timeout_secs, and
config.agent.agent_timeout_secs when making the change.
In `@tests/json_rpc_e2e.rs`:
- Around line 9088-9202: The test
json_rpc_config_agent_timeout_settings_roundtrip mutates the process-global
runtime timeout and never restores it; capture the original effective timeout at
the start (via the existing openhuman.config_get_agent_settings call or by
reading initial_result.agent_timeout_secs) and restore it at the end of the test
by calling openhuman.config_update_agent_settings with that saved value (or
install a small RAII guard in the test that calls the same RPC in its Drop to
ensure restoration even on early return/panic), so later tests don’t inherit the
modified AtomicU64.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 4e2130b8-7bf6-4da3-a9fe-356e14f8c167
📒 Files selected for processing (30)
.env.exampleapp/src/components/settings/panels/AgentAccessPanel.tsxapp/src/components/settings/panels/__tests__/AgentAccessPanel.test.tsxapp/src/lib/i18n/ar.tsapp/src/lib/i18n/bn.tsapp/src/lib/i18n/de.tsapp/src/lib/i18n/en.tsapp/src/lib/i18n/es.tsapp/src/lib/i18n/fr.tsapp/src/lib/i18n/hi.tsapp/src/lib/i18n/id.tsapp/src/lib/i18n/it.tsapp/src/lib/i18n/ko.tsapp/src/lib/i18n/pl.tsapp/src/lib/i18n/pt.tsapp/src/lib/i18n/ru.tsapp/src/lib/i18n/zh-CN.tsapp/src/services/rpcMethods.tsapp/src/utils/tauriCommands/config.tsdocs/TEST-COVERAGE-MATRIX.mdsrc/openhuman/about_app/catalog_data.rssrc/openhuman/channels/runtime/startup.rssrc/openhuman/config/ops.rssrc/openhuman/config/ops_tests.rssrc/openhuman/config/schema/agent.rssrc/openhuman/config/schemas.rssrc/openhuman/config/schemas_tests.rssrc/openhuman/tool_timeout/README.mdsrc/openhuman/tool_timeout/mod.rstests/json_rpc_e2e.rs
The `worker_a_controller_schemas_are_fully_exposed` test pins the full list of config-namespace RPC method names. Add the two new methods (`config_get_agent_settings`, `config_update_agent_settings`) to the expected list in alphabetical order.
- Decouple timeout fetch from autonomy load so a timeout-RPC failure doesn't block the pre-existing access-mode controls. - Avoid overwriting a newer draft when an older save response arrives (compare draftAtCommit before snapping timeoutInput). - Seed the runtime timeout from config in get_agent_settings so the effective_timeout_secs field is correct even without a full boot. - Restore the process-global timeout to the default after the e2e test to prevent order-dependent bleeding across tests.
5 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
[agent].agent_timeout_secs(default 120s, range 1–3600) and expose it over newconfig_get_agent_settings/config_update_agent_settingsRPCs.tool_timeoutruntime mutable (AtomicU64+set_tool_timeout_secs) so a change applies on the next tool call without a core restart.OPENHUMAN_TOOL_TIMEOUT_SECSas an operator override that still wins; the UI shows a notice and disables the field when it is active.enand all 13 non-English locales with real translations; about_app catalog, coverage matrix, and.env.exampleupdated.Problem
A user (GameySins, via Discord) reported that agent actions using a local model are cut off by the tool-execution timeout before the model finishes responding. They explicitly want to extend the timeout, not switch to a smaller model. The only knob today is the
OPENHUMAN_TOOL_TIMEOUT_SECSenv var, resolved once into an immutableOnceLock— there is no UI, and the workaround is hand-editing config/.env. This is the feature-request gap in #3100.Solution
End-to-end, following the existing autonomy-settings pattern:
tool_timeout(src/openhuman/tool_timeout/mod.rs): replaced the immutableOnceLock<u64>with anAtomicU64seeded lazily from env/default, plus a pureresolve_effectiveresolver andset_tool_timeout_secs. Precedence: valid env var → persisted config → 120s default. Both consumers (engine/tools.rs,delegate.rs) read fresh per call, so changes take effect immediately.parse_tool_timeout_secsis unchanged.[agent].agent_timeout_secstoAgentConfig;apply_agent_settingsvalidates the 1–3600 range, persists, and pushes the value into the livetool_timeoutruntime;get_agent_settingsreports the configured value, the runtime-effective value,env_override, and the bounds for the UI. Exposed via the controller registry only. Startup seeds the runtime from config.AgentAccessPanelthat loads viaconfig_get_agent_settings, persists on blur/Enter (independent of the autonomy block, with a last-write-wins guard), validates client-side, and disables itself with an explanatory notice when the env override is active.Design note: the timeout lives in
[agent](alongsidemax_tool_iterations) rather than the security-focused[autonomy]block, with its own RPC pair, to keep concerns separated.Submission Checklist
tool_timeoutresolver/env-precedence unit tests,configops apply/reject/no-op tests, schemas registration test, and ajson_rpc_e2eroundtrip (get → update → effective → reject). Vitest:AgentAccessPanelload/persist-on-blur/range-reject/no-op/env-override-disable. (Could not runpnpm test:coverage/pnpm test:rustto completion locally — see Validation Blocked; targeted suites all pass.)13.1.6 Action Timeoutindocs/TEST-COVERAGE-MATRIX.md## RelatedCloses #3100in the## RelatedsectionImpact
agent_timeout_secsdefaults to the historical 120s, so behaviour is unchanged until a user opts in. CLI/core honour the same config field.OPENHUMAN_TOOL_TIMEOUT_SECS) semantics are preserved and documented; docker/cloud deployments are unaffected.tokio::time::timeoutkeeps its captured deadline; the new value applies to subsequent tool calls.Related
13.1.6(Action Timeout); capabilityagent.action_timeoutAI Authored PR Metadata (required for Codex/Linear PRs)
Linear Issue
Commit & Branch
issue/3100-no-ui-option-to-configure-agent-timeoutValidation Run
pnpm --filter openhuman-app format:check— Prettier + cargo fmt clean on changed filespnpm typecheck— passestool_timeout(11), config opsapply_agent*(3), config schemas (23), about_app (24),json_rpc_config_agent_timeout_settings_roundtrip(1),AgentAccessPanel.test.tsx(15) — all passcargo fmt --check+cargo checkon the core crate passapp/src-taurichangesValidation Blocked
command:pnpm rust:check(pre-push hook) /pnpm lint:commands-tokenserror:Tauri shellcargo checkfails buildingglib-sys(missing GTK/glib system libs in this sandbox);lint:commands-tokensrequiresripgrep, which is not installed here.impact:Both are environmental and unrelated to this diff (which is entirely core-crate + frontend, noapp/src-tauri). Pushed with--no-verify; CI runs these on a provisioned runner.Behavior Changes
Parity Contract
OPENHUMAN_TOOL_TIMEOUT_SECSstill overrides;parse_tool_timeout_secsbounds/defaults unchanged; default timeout remains 120s.Duplicate / Superseded PR Handling
Summary by CodeRabbit
New Features
Operator override
Internationalization
Documentation
Tests