admin/openhuman
1
0
Fork 0
mirror of https://github.com/tinyhumansai/openhuman.git synced 2026-07-27 21:08:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat(agent): steer_subagent + wait_subagent for running async sub-agents (#3641)

Browse Source
This commit is contained in:
Steven Enamakel
2026-06-12 17:02:32 -07:00
committed by GitHub
parent abddfcbd1b
commit 75a6f2891b
20 changed files with 1084 additions and 175 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/TEST-COVERAGE-MATRIX.md
+180 -172
View File
@@ -46,12 +46,12 @@ Canonical mapping of every product feature to its test source(s). Drives gap-fil
### 0.3 Updates & Reinstallation
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ----------------------------- | ----- | -------------------------------------------------- | ------ | ------------------------------------- |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ----------------------------- | -------- | --------------------------------------------------------------------------- | ------ | ------------------------------------------------------------------------------- |
| 0.3.1 | Auto Update Check | RU+RI+MS | `src/openhuman/update/` (Rust unit), `tests/json_rpc_e2e.rs`, release smoke | 🟡 | Core check/update policy covered; desktop prompt + release upgrade still manual |
| 0.3.2 | Forced Update Handling | MS | release-manual-smoke | 🚫 | End-to-end gating verified at release |
| 0.3.3 | Reinstall with Existing State | MS | release-manual-smoke | 🚫 | Workspace persistence on reinstall |
| 0.3.4 | Clean Uninstall | MS | release-manual-smoke | 🚫 | OS removal paths |
| 0.3.2 | Forced Update Handling | MS | release-manual-smoke | 🚫 | End-to-end gating verified at release |
| 0.3.3 | Reinstall with Existing State | MS | release-manual-smoke | 🚫 | Workspace persistence on reinstall |
| 0.3.4 | Clean Uninstall | MS | release-manual-smoke | 🚫 | OS removal paths |
---
@@ -118,22 +118,22 @@ Canonical mapping of every product feature to its test source(s). Drives gap-fil
### 3.1 Model Management
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ----------------------------- | ----- | -------------------------------------------------------- | ------ | ----- |
| 3.1.1 | Model Detection | RU+WD | `src/openhuman/local_ai/`, `local-model-runtime.spec.ts` | ✅ | |
| 3.1.2 | Model Download & Installation | WD | `local-model-runtime.spec.ts` | ✅ | |
| 3.1.3 | Model Version Handling | RU | `src/openhuman/local_ai/model_ids.rs` | ✅ | |
| 3.1.4 | LM Studio Model Discovery | RU+RI | `src/openhuman/local_ai/service/ollama_admin_tests.rs`, `tests/json_rpc_e2e.rs` | ✅ | Uses LM Studio's OpenAI-compatible `/v1/models` surface |
| 3.1.5 | Model Context-Window Requirement Gate | RU+VU | `src/openhuman/inference/local/model_requirements.rs`, `src/openhuman/inference/local/ollama.rs`, `src/openhuman/inference/local/service/ollama_admin_tests.rs`, `app/src/components/settings/panels/local-model/ModelStatusSection.test.tsx` | ✅ | Rejects Ollama models whose native context window is below the memory-layer minimum (`local_ai.model_context_check`) |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ------------------------------------- | ----- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------ | -------------------------------------------------------------------------------------------------------------------- |
| 3.1.1 | Model Detection | RU+WD | `src/openhuman/local_ai/`, `local-model-runtime.spec.ts` | ✅ | |
| 3.1.2 | Model Download & Installation | WD | `local-model-runtime.spec.ts` | ✅ | |
| 3.1.3 | Model Version Handling | RU | `src/openhuman/local_ai/model_ids.rs` | ✅ | |
| 3.1.4 | LM Studio Model Discovery | RU+RI | `src/openhuman/local_ai/service/ollama_admin_tests.rs`, `tests/json_rpc_e2e.rs` | ✅ | Uses LM Studio's OpenAI-compatible `/v1/models` surface |
| 3.1.5 | Model Context-Window Requirement Gate | RU+VU | `src/openhuman/inference/local/model_requirements.rs`, `src/openhuman/inference/local/ollama.rs`, `src/openhuman/inference/local/service/ollama_admin_tests.rs`, `app/src/components/settings/panels/local-model/ModelStatusSection.test.tsx` | ✅ | Rejects Ollama models whose native context window is below the memory-layer minimum (`local_ai.model_context_check`) |
### 3.2 Runtime Execution
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ---------------------------------- | ----- | ---------------------------------- | ------ | ----------------------------------------- |
| 3.2.1 | Local Inference Execution | WD | `local-model-runtime.spec.ts` | ✅ | |
| 3.2.2 | Resource Handling (CPU/GPU/Memory) | RU | `src/openhuman/local_ai/device.rs` | 🟡 | Detection unit; runtime constraint manual |
| 3.2.3 | Runtime Failure Handling | RU+WD | `local-model-runtime.spec.ts` | ✅ | |
| 3.2.4 | LM Studio Chat Completions | RU+RI | `src/openhuman/local_ai/service/public_infer_tests.rs`, `tests/json_rpc_e2e.rs` | ✅ | Covers prompt/chat success and non-success status errors |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ---------------------------------- | ----- | ------------------------------------------------------------------------------- | ------ | -------------------------------------------------------- |
| 3.2.1 | Local Inference Execution | WD | `local-model-runtime.spec.ts` | ✅ | |
| 3.2.2 | Resource Handling (CPU/GPU/Memory) | RU | `src/openhuman/local_ai/device.rs` | 🟡 | Detection unit; runtime constraint manual |
| 3.2.3 | Runtime Failure Handling | RU+WD | `local-model-runtime.spec.ts` | ✅ | |
| 3.2.4 | LM Studio Chat Completions | RU+RI | `src/openhuman/local_ai/service/public_infer_tests.rs`, `tests/json_rpc_e2e.rs` | ✅ | Covers prompt/chat success and non-success status errors |
### 3.3 Runtime Configuration
@@ -155,12 +155,12 @@ Canonical mapping of every product feature to its test source(s). Drives gap-fil
#### 3.3.3 Configuration Persistence
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------- | ----------------- | ----- | ----------------------------- | ------ | --------------------- |
| 3.3.3.1 | Save RAM Settings | VU | _missing_ | ❌ | Settings slice |
| 3.3.3.2 | Apply on Restart | WD | `local-model-runtime.spec.ts` | 🟡 | Restart not exercised |
| 3.3.3.3 | Reset to Default | VU | _missing_ | ❌ | |
| 3.3.3.4 | Provider Selection Persistence | RU+RI+VU | `src/openhuman/config/ops_tests.rs`, `tests/json_rpc_e2e.rs`, `app/src/utils/tauriCommands/config.test.ts` | ✅ | Covers `lm_studio` normalization and config round-trip |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------- | ------------------------------ | -------- | ---------------------------------------------------------------------------------------------------------- | ------ | ------------------------------------------------------ |
| 3.3.3.1 | Save RAM Settings | VU | _missing_ | ❌ | Settings slice |
| 3.3.3.2 | Apply on Restart | WD | `local-model-runtime.spec.ts` | 🟡 | Restart not exercised |
| 3.3.3.3 | Reset to Default | VU | _missing_ | ❌ | |
| 3.3.3.4 | Provider Selection Persistence | RU+RI+VU | `src/openhuman/config/ops_tests.rs`, `tests/json_rpc_e2e.rs`, `app/src/utils/tauriCommands/config.test.ts` | ✅ | Covers `lm_studio` normalization and config round-trip |
---
@@ -176,24 +176,24 @@ Canonical mapping of every product feature to its test source(s). Drives gap-fil
### 4.2 Messaging
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ---------------------- | ----- | ----------------------------------------------------------------- | ------ | --------------------------- |
| 4.2.1 | User Message Handling | WD+RI | `conversations-web-channel-flow.spec.ts`, `tests/json_rpc_e2e.rs` | ✅ | |
| 4.2.2 | AI Response Generation | WD | `agent-review.spec.ts` | ✅ | Mock LLM |
| 4.2.3 | Streaming Responses | RI | `tests/json_rpc_e2e.rs` | 🟡 | UI streaming assertion thin |
| 4.2.4 | Parallel inference (cross-thread + within-thread forked turns) | RU+VU | `src/openhuman/channels/providers/web_tests.rs`, `app/src/store/__tests__/chatRuntimeSlice.test.ts`, `app/src/providers/__tests__/ChatRuntimeProvider.test.tsx` | 🟡 | Concurrent same-/cross-thread dispatch, cooperative `CancellationToken` teardown, and parallel-lane stream routing covered; dedicated WD E2E is a follow-up |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | -------------------------------------------------------------- | ----- | --------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------- |
| 4.2.1 | User Message Handling | WD+RI | `conversations-web-channel-flow.spec.ts`, `tests/json_rpc_e2e.rs` | ✅ | |
| 4.2.2 | AI Response Generation | WD | `agent-review.spec.ts` | ✅ | Mock LLM |
| 4.2.3 | Streaming Responses | RI | `tests/json_rpc_e2e.rs` | 🟡 | UI streaming assertion thin |
| 4.2.4 | Parallel inference (cross-thread + within-thread forked turns) | RU+VU | `src/openhuman/channels/providers/web_tests.rs`, `app/src/store/__tests__/chatRuntimeSlice.test.ts`, `app/src/providers/__tests__/ChatRuntimeProvider.test.tsx` | 🟡 | Concurrent same-/cross-thread dispatch, cooperative `CancellationToken` teardown, and parallel-lane stream routing covered; dedicated WD E2E is a follow-up |
### 4.3 Tool Invocation
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | -------------------------- | ----- | ----------------------------------------------------------- | ------ | ----- |
| 4.3.1 | Tool Trigger via Chat | WD | `skill-execution-flow.spec.ts`, `skill-multi-round.spec.ts` | ✅ | |
| 4.3.2 | Permission-Based Execution | RU+WD | `src/openhuman/tools/`, `skill-execution-flow.spec.ts` | ✅ | |
| 4.3.3 | Tool Failure Handling | WD | `skill-execution-flow.spec.ts` | ✅ | |
| 4.3.4 | Subagent Mascot Visualization | VU | `app/src/features/human/SubMascotLayer.test.tsx`, `app/src/features/human/HumanPage.test.tsx` | ✅ | Renders spawned/completed/failed subagent timeline rows as colored companion mascots with activity bubbles |
| 4.3.5 | Image Tool Contracts | RU | `src/openhuman/image/` | ✅ | High-level `image_generation` / `view_image` schema, gating, serialization, prompt guidance, and contract e2e coverage for #2984 |
| 4.3.6 | Background Monitor Tools | RU+RI | `src/openhuman/monitor/`, `src/openhuman/tools/ops_tests.rs`, `tests/json_rpc_e2e.rs` | ✅ | First-class monitor domain covers command denial, line streaming, timeout, stop, bounded output, registry exposure, and JSON-RPC list/read surface for #3371 |
| 4.3.7 | Mascot Avatar Animation | VU | `app/src/features/human/Mascot/RiveMascot.test.tsx`, `app/src/features/human/Mascot/riveMaps.test.ts` | ✅ | Rive `MascotSM` state machine: face→pose mapping, Oculus→`visme_codes` viseme normalization, and idle random pose rotation for the `tiny_mascot.riv` upgrade |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ----------------------------- | ----- | ----------------------------------------------------------------------------------------------------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| 4.3.1 | Tool Trigger via Chat | WD | `skill-execution-flow.spec.ts`, `skill-multi-round.spec.ts` | ✅ | |
| 4.3.2 | Permission-Based Execution | RU+WD | `src/openhuman/tools/`, `skill-execution-flow.spec.ts` | ✅ | |
| 4.3.3 | Tool Failure Handling | WD | `skill-execution-flow.spec.ts` | ✅ | |
| 4.3.4 | Subagent Mascot Visualization | VU | `app/src/features/human/SubMascotLayer.test.tsx`, `app/src/features/human/HumanPage.test.tsx` | ✅ | Renders spawned/completed/failed subagent timeline rows as colored companion mascots with activity bubbles |
| 4.3.5 | Image Tool Contracts | RU | `src/openhuman/image/` | ✅ | High-level `image_generation` / `view_image` schema, gating, serialization, prompt guidance, and contract e2e coverage for #2984 |
| 4.3.6 | Background Monitor Tools | RU+RI | `src/openhuman/monitor/`, `src/openhuman/tools/ops_tests.rs`, `tests/json_rpc_e2e.rs` | ✅ | First-class monitor domain covers command denial, line streaming, timeout, stop, bounded output, registry exposure, and JSON-RPC list/read surface for #3371 |
| 4.3.7 | Mascot Avatar Animation | VU | `app/src/features/human/Mascot/RiveMascot.test.tsx`, `app/src/features/human/Mascot/riveMaps.test.ts` | ✅ | Rive `MascotSM` state machine: face→pose mapping, Oculus→`visme_codes` viseme normalization, and idle random pose rotation for the `tiny_mascot.riv` upgrade |
---
@@ -201,36 +201,36 @@ Canonical mapping of every product feature to its test source(s). Drives gap-fil
### 5.1 Screen Intelligence
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ------------------ | ----- | ------------------------------------------------------------------------ | ------ | ----- |
| 5.1.1 | Screen Capture | RI | `tests/screen_intelligence_vision_e2e.rs` | ✅ | |
| 5.1.2 | Context Extraction | RI | `tests/screen_intelligence_vision_e2e.rs` | ✅ | |
| 5.1.3 | Memory Injection | RI | `tests/memory_graph_sync_e2e.rs` | ✅ | |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ------------------ | ----- | ----------------------------------------- | ------ | ----- |
| 5.1.1 | Screen Capture | RI | `tests/screen_intelligence_vision_e2e.rs` | ✅ | |
| 5.1.2 | Context Extraction | RI | `tests/screen_intelligence_vision_e2e.rs` | ✅ | |
| 5.1.3 | Memory Injection | RI | `tests/memory_graph_sync_e2e.rs` | ✅ | |
### 5.2 Text Autocomplete
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ---------------------------- | ----- | ------------------------------------------------------------------------------------------------------------------------------------------- | ------ | ----------------------------------------------------------------------------------- |
| 5.2.1 | Inline Suggestion Generation | MS+WD | `app/test/e2e/specs/autocomplete-flow.spec.ts` (settings surface only); release-manual-smoke for real inline-gen | 🟡 | Settings panel mounts; inline-gen requires macOS TCC grants — manual only |
| 5.2.2 | Debounce Handling | VU | `app/src/features/autocomplete/__tests__/useAutocompleteSkillStatus.test.tsx` (status surface); core debounce timing is Rust-side | ✅ | Was ❌ — status branches now covered |
| 5.2.3 | Acceptance Trigger | MS | release-manual-smoke (#971) | 🟡 | Real keypress acceptance into a third-party text field — not driver-automatable |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ---------------------------- | ----- | --------------------------------------------------------------------------------------------------------------------------------- | ------ | ------------------------------------------------------------------------------- |
| 5.2.1 | Inline Suggestion Generation | MS+WD | `app/test/e2e/specs/autocomplete-flow.spec.ts` (settings surface only); release-manual-smoke for real inline-gen | 🟡 | Settings panel mounts; inline-gen requires macOS TCC grants — manual only |
| 5.2.2 | Debounce Handling | VU | `app/src/features/autocomplete/__tests__/useAutocompleteSkillStatus.test.tsx` (status surface); core debounce timing is Rust-side | ✅ | Was ❌ — status branches now covered |
| 5.2.3 | Acceptance Trigger | MS | release-manual-smoke (#971) | 🟡 | Real keypress acceptance into a third-party text field — not driver-automatable |
### 5.3 Voice Intelligence
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ------------------------- | ----- | -------------------- | ------ | ----- |
| 5.3.1 | Voice Input Capture | WD | `voice-mode.spec.ts` | ✅ | |
| 5.3.2 | Speech-to-Text Processing | WD | `voice-mode.spec.ts` | ✅ | |
| 5.3.3 | Voice Command Execution | WD | `voice-mode.spec.ts` | ✅ | |
| 5.3.4 | Mascot Voice Selection | VU | `app/src/store/__tests__/mascotSlice.test.ts`, `app/src/components/settings/panels/__tests__/VoicePanel.test.tsx`, `app/src/features/human/useHumanMascot.test.ts` | ✅ | Slice validation + persist REHYDRATE, Settings picker UI (#1762), `synthesizeSpeech` voiceId override propagation |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ------------------------- | ----- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------ | ----------------------------------------------------------------------------------------------------------------- |
| 5.3.1 | Voice Input Capture | WD | `voice-mode.spec.ts` | ✅ | |
| 5.3.2 | Speech-to-Text Processing | WD | `voice-mode.spec.ts` | ✅ | |
| 5.3.3 | Voice Command Execution | WD | `voice-mode.spec.ts` | ✅ | |
| 5.3.4 | Mascot Voice Selection | VU | `app/src/store/__tests__/mascotSlice.test.ts`, `app/src/components/settings/panels/__tests__/VoicePanel.test.tsx`, `app/src/features/human/useHumanMascot.test.ts` | ✅ | Slice validation + persist REHYDRATE, Settings picker UI (#1762), `synthesizeSpeech` voiceId override propagation |
### 5.4 Persona
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ----------------------------- | ----- | ----------------------------------------------------------------------------------------------------------------------------------- | ------ | ------------------------------------------------------------------------------------------------- |
| 5.4.1 | Persona Name & Description | VU | `app/src/store/personaSlice.test.ts`, `app/src/components/settings/panels/PersonaPanel.test.tsx` | ✅ | Slice validation + persist REHYDRATE scrub; Settings identity fields persist on save (#2345) |
| 5.4.2 | SOUL.md Edit & Reset | RU+VU | `src/openhuman/workspace/rpc.rs`, `app/src/components/settings/panels/PersonaPanel.test.tsx` | ✅ | Core read/write/reset with allowlist + size cap; panel loads, saves, resets over RPC (#2345) |
| 5.4.3 | Persona Settings Surface | VU | `app/src/components/settings/panels/PersonaPanel.test.tsx` | ✅ | Bundles identity + SOUL.md + link to Mascot avatar/voice (#2345) |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | -------------------------- | ----- | ------------------------------------------------------------------------------------------------ | ------ | -------------------------------------------------------------------------------------------- |
| 5.4.1 | Persona Name & Description | VU | `app/src/store/personaSlice.test.ts`, `app/src/components/settings/panels/PersonaPanel.test.tsx` | ✅ | Slice validation + persist REHYDRATE scrub; Settings identity fields persist on save (#2345) |
| 5.4.2 | SOUL.md Edit & Reset | RU+VU | `src/openhuman/workspace/rpc.rs`, `app/src/components/settings/panels/PersonaPanel.test.tsx` | ✅ | Core read/write/reset with allowlist + size cap; panel loads, saves, resets over RPC (#2345) |
| 5.4.3 | Persona Settings Surface | VU | `app/src/components/settings/panels/PersonaPanel.test.tsx` | ✅ | Bundles identity + SOUL.md + link to Mascot avatar/voice (#2345) |
---
@@ -238,39 +238,47 @@ Canonical mapping of every product feature to its test source(s). Drives gap-fil
### 6.1 File System
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ---------------------------- | ----- | ---------------------------------------------------------------------------------------------------------------- | ------ | -------------------------------------------------------------------- |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ---------------------------- | ----- | ------------------------------------------------------------------------------------------------------ | ------ | -------------------------------------------------------------------- |
| 6.1.1 | File Read Access | RU+WD | `src/openhuman/tools/impl/filesystem/file_read.rs`, `app/test/e2e/specs/tool-filesystem-flow.spec.ts` | ✅ | Was 🟡 — WDIO drives memory_read_file + asserts via Node fs |
| 6.1.2 | File Write Access | RU+WD | `src/openhuman/tools/impl/filesystem/file_write.rs`, `app/test/e2e/specs/tool-filesystem-flow.spec.ts` | ✅ | Was 🟡 — WDIO drives memory_write_file + asserts bytes match on disk |
| 6.1.3 | Path Restriction Enforcement | RU+WD | `src/openhuman/tools/impl/filesystem/file_read.rs`, `app/test/e2e/specs/tool-filesystem-flow.spec.ts` | ✅ | Was 🟡 — WDIO asserts traversal + absolute-path denial envelope |
### 6.2 Shell & Git
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ---------------------------- | ----- | ------------------------------------------------------------------------------------------------------------------------- | ------ | ------------------------------------------------------------------------------------------------ |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ---------------------------- | ----- | --------------------------------------------------------------------------------------------------------------- | ------ | ------------------------------------------------------------------------------------------------ |
| 6.2.1 | Shell Command Execution | RU+WD | `src/openhuman/tools/impl/system/shell.rs`, `app/test/e2e/specs/tool-shell-git-flow.spec.ts` | ✅ | Was 🟡 — WDIO asserts agent runtime + `tools_agent` registry contract; full LLM path tracked #68 |
| 6.2.2 | Command Restriction Handling | RU+WD | `src/openhuman/security/policy_tests.rs`, `app/test/e2e/specs/tool-shell-git-flow.spec.ts` | ✅ | Was 🟡 — WDIO locks denial envelope shape `{ ok:false, error }` consumed by the React UI |
| 6.2.3 | Git Read Operations | RU+WD | `src/openhuman/tools/impl/filesystem/git_operations_tests.rs`, `app/test/e2e/specs/tool-shell-git-flow.spec.ts` | ✅ | Was 🟡 — WDIO seeds a fixture repo in OPENHUMAN_WORKSPACE and asserts read ops succeed |
| 6.2.4 | Git Write Operations | RU+WD | `src/openhuman/tools/impl/filesystem/git_operations_tests.rs`, `app/test/e2e/specs/tool-shell-git-flow.spec.ts` | ✅ | Was 🟡 — WDIO commits into the same fixture and asserts log advances |
### 6.3 Sub-agent Orchestration
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ---------------------------- | ----- | --------------------------------------------------------------------------------------------------------------------- | ------ | -------------------------------------------------------------------------------------------------------------------------------------------------- |
| 6.3.1 | Steer a running sub-agent | RU | `src/openhuman/agent_orchestration/running_subagents.rs`, `src/openhuman/agent_orchestration/tools/steer_subagent.rs` | ✅ | `steer_subagent` injects a steer/collect message into a running async sub-agent's run-queue; registry enforces parent ownership + terminal guard. |
| 6.3.2 | Wait for a sub-agent result | RU | `src/openhuman/agent_orchestration/running_subagents.rs`, `src/openhuman/agent_orchestration/tools/wait_subagent.rs` | ✅ | `wait_subagent` blocks on the completion `watch` with a timeout; prunes terminal entries, leaves entries intact on timeout. |
| 6.3.3 | Steer lands in child history | RU | `src/openhuman/agent/harness/subagent_runner/ops_tests.rs::run_queue_steer_lands_in_subagent_history` | ✅ | End-to-end: a queued steer is drained by the child `run_turn_engine` and appears as a `[User steering message]` user turn in the provider request. |
---
## 7. Web & Network Capabilities
### 7.1 Browser
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ------------------ | ----- | ------------------------------------------------------------------------------------------------------------------ | ------ | --------------------------------------------------------------------------------------------------- |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ------------------ | ----- | -------------------------------------------------------------------------------------------------------- | ------ | --------------------------------------------------------------------------------------------------- |
| 7.1.1 | Open URL | RU+WD | `src/openhuman/tools/impl/browser/browser_open_tests.rs`, `app/test/e2e/specs/tool-browser-flow.spec.ts` | ✅ | Was ❌ — WDIO asserts agent runtime + browser-bearing registry; mock backend captures HTTP shape |
| 7.1.2 | Browser Automation | RU+WD | `src/openhuman/tools/impl/browser/browser_tests.rs`, `app/test/e2e/specs/tool-browser-flow.spec.ts` | ✅ | Was ❌ — WDIO locks tools_agent wildcard scope (exposes the 22-action automation schema to the LLM) |
### 7.2 Network
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | -------------------- | ----- | ----------------------------------- | ------ | ------------------ |
| 7.2.1 | HTTP / API Requests | RU+WD | `service-connectivity-flow.spec.ts` | ✅ | |
| 7.2.2 | Web Search Execution | WD | `skill-execution-flow.spec.ts` | 🟡 | Generic skill path |
| 7.2.3 | TinyFish Integration Tools | RU | `src/openhuman/integrations/tinyfish_tests.rs`, `src/openhuman/tools/ops_tests.rs::all_tools_executes_tinyfish_family_against_fake_backend` | ✅ | Backend-proxied Search, Fetch, and Agent run tools covered with fake backend |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | -------------------------- | ----- | ------------------------------------------------------------------------------------------------------------------------------------------- | ------ | ---------------------------------------------------------------------------- |
| 7.2.1 | HTTP / API Requests | RU+WD | `service-connectivity-flow.spec.ts` | ✅ | |
| 7.2.2 | Web Search Execution | WD | `skill-execution-flow.spec.ts` | 🟡 | Generic skill path |
| 7.2.3 | TinyFish Integration Tools | RU | `src/openhuman/integrations/tinyfish_tests.rs`, `src/openhuman/tools/ops_tests.rs::all_tools_executes_tinyfish_family_against_fake_backend` | ✅ | Backend-proxied Search, Fetch, and Agent run tools covered with fake backend |
---
@@ -278,41 +286,41 @@ Canonical mapping of every product feature to its test source(s). Drives gap-fil
### 8.1 Memory Operations
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ------------- | ----- | -------------------------------------------------------------------------------------------------- | ------ | ------ |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ------------- | ----- | ------------------------------------------------------------------------------ | ------ | ------ |
| 8.1.1 | Store Memory | RI+WD | `tests/memory_roundtrip_e2e.rs`, `app/test/e2e/specs/memory-roundtrip.spec.ts` | ✅ | Was ❌ |
| 8.1.2 | Recall Memory | RI+WD | same | ✅ | Was ❌ |
| 8.1.3 | Forget Memory | RI+WD | same | ✅ | Was ❌ |
| 8.1.2 | Recall Memory | RI+WD | same | ✅ | Was ❌ |
| 8.1.3 | Forget Memory | RI+WD | same | ✅ | Was ❌ |
### 8.2 Memory Handling
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ------------------ | ----- | ----------------------------------------- | ------ | --------------------------------- |
| 8.2.1 | Context Injection | RI | `tests/autocomplete_memory_e2e.rs` | ✅ | |
| 8.2.2 | Memory Consistency | RI | `tests/memory_graph_sync_e2e.rs`, `tests/worker_c_modules_e2e.rs` | ✅ | Worker C RPC E2E verifies memory-tree ingest is reflected by `memory_sync_status_list` |
| 8.2.3 | Memory Scaling | RU | `src/openhuman/memory/ingestion_tests.rs` | 🟡 | Soak/scale benchmark not asserted |
| 8.2.4 | Raw-archive sync reconcile | RU+RI | `src/openhuman/memory_sync/sources/rebuild.rs`, `src/openhuman/memory_sync/workspace/periodic.rs`, `tests/json_rpc_e2e.rs` (`json_rpc_memory_sources_reconcile_reports_pending_raw_files`), `tests/memory_sync_pipeline_e2e.rs` | ✅ | Coverage gate + incremental rebuild + workspace periodic scheduler + `memory_sources_reconcile` RPC |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | -------------------------- | ----- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------ | --------------------------------------------------------------------------------------------------- |
| 8.2.1 | Context Injection | RI | `tests/autocomplete_memory_e2e.rs` | ✅ | |
| 8.2.2 | Memory Consistency | RI | `tests/memory_graph_sync_e2e.rs`, `tests/worker_c_modules_e2e.rs` | ✅ | Worker C RPC E2E verifies memory-tree ingest is reflected by `memory_sync_status_list` |
| 8.2.3 | Memory Scaling | RU | `src/openhuman/memory/ingestion_tests.rs` | 🟡 | Soak/scale benchmark not asserted |
| 8.2.4 | Raw-archive sync reconcile | RU+RI | `src/openhuman/memory_sync/sources/rebuild.rs`, `src/openhuman/memory_sync/workspace/periodic.rs`, `tests/json_rpc_e2e.rs` (`json_rpc_memory_sources_reconcile_reports_pending_raw_files`), `tests/memory_sync_pipeline_e2e.rs` | ✅ | Coverage gate + incremental rebuild + workspace periodic scheduler + `memory_sources_reconcile` RPC |
### 8.3 Memory Retrieval Benchmarks
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ---------------------------------------- | ----- | ---------------------------------------------------------------------------------- | ------ | ----- |
| 8.3.2 | Cross-Chat Entity Discoverability | RU | `src/openhuman/memory/tree/retrieval/benchmarks.rs::bench_cross_chat_entity_discoverable` | ✅ | Verifies entity canonicalisation across multiple chats |
| 8.3.3 | Citation Bundle Provenance | RU | `src/openhuman/memory/tree/retrieval/benchmarks.rs::bench_citation_bundle_provenance` | ✅ | Verifies source_ref and tree_scope are populated in retrieval hits |
| 8.3.4 | Citation Fetch Leaves Hydration | RU | `src/openhuman/memory/tree/retrieval/benchmarks.rs::bench_citation_fetch_leaves_hydrates` | ✅ | Verifies fetch_leaves returns content for exact chunk IDs |
| 8.3.7 | Long-Source Exact Leaf Retrieval | RU | `src/openhuman/memory/tree/retrieval/benchmarks.rs::bench_long_source_retrieves_exact_leaf` | 🟡 | Embedder required for seal + chunking; test runs in inert mode but assertions are conditional |
| 8.3.9 | Scale Ingest 20 Sources No Real Data | RU | `src/openhuman/memory/tree/retrieval/benchmarks.rs::bench_scale_ingest_20_sources_no_real_data` | ✅ | Verifies retrieval correctness at scale with synthetic data |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ------------------------------------ | ----- | ----------------------------------------------------------------------------------------------- | ------ | --------------------------------------------------------------------------------------------- |
| 8.3.2 | Cross-Chat Entity Discoverability | RU | `src/openhuman/memory/tree/retrieval/benchmarks.rs::bench_cross_chat_entity_discoverable` | ✅ | Verifies entity canonicalisation across multiple chats |
| 8.3.3 | Citation Bundle Provenance | RU | `src/openhuman/memory/tree/retrieval/benchmarks.rs::bench_citation_bundle_provenance` | ✅ | Verifies source_ref and tree_scope are populated in retrieval hits |
| 8.3.4 | Citation Fetch Leaves Hydration | RU | `src/openhuman/memory/tree/retrieval/benchmarks.rs::bench_citation_fetch_leaves_hydrates` | ✅ | Verifies fetch_leaves returns content for exact chunk IDs |
| 8.3.7 | Long-Source Exact Leaf Retrieval | RU | `src/openhuman/memory/tree/retrieval/benchmarks.rs::bench_long_source_retrieves_exact_leaf` | 🟡 | Embedder required for seal + chunking; test runs in inert mode but assertions are conditional |
| 8.3.9 | Scale Ingest 20 Sources No Real Data | RU | `src/openhuman/memory/tree/retrieval/benchmarks.rs::bench_scale_ingest_20_sources_no_real_data` | ✅ | Verifies retrieval correctness at scale with synthetic data |
### 8.4 Explicit User Preferences (Two-Lane)
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ------------------------------------------ | ----- | ----------------------------------------------------------------------------------------------------------------- | ------ | ---------------------------------------------------------------------- |
| 8.4.1 | Save Preference (general / situational) | RU | `src/openhuman/agent/tools/save_preference_tests.rs` | ✅ | `save_preference` tool → `user_pref_{general,situational}`, topic-keyed |
| 8.4.2 | Lane A — Standing Prefs in System Prompt | RU | `src/openhuman/learning/prompt_sections.rs`, `src/openhuman/agent/harness/session/turn_tests.rs` | ✅ | General prefs rendered into the system prompt at thread start |
| 8.4.3 | Lane B — Situational Recall (vector-gated) | RU | `src/openhuman/memory/store/unified/query_tests.rs::recall_relevant_by_vector_gates_on_similarity` | ✅ | Per-turn; relevant query injects, unrelated suppresses |
| 8.4.4 | Same-Topic Contradiction (replace) | RU | `src/openhuman/agent/tools/save_preference_tests.rs::recategorising_moves_pref_between_namespaces` | ✅ | `ON CONFLICT REPLACE`; a topic lives in exactly one scope |
| 8.4.5 | Cross-Topic Contradiction Surfacing | RU | `src/openhuman/agent/tools/save_preference_tests.rs::save_surfaces_related_preference_for_contradiction_check` | ✅ | Related prefs surfaced in the tool result for the chat agent to resolve |
| 8.4.6 | vector_chunks Model-Signature Recall Guard | RU | `src/openhuman/memory/store/unified/query_tests.rs::vector_recall_excludes_other_model_signature` | ✅ | Excludes cross-model vectors; dim-guards legacy rows |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ----- | ------------------------------------------ | ----- | -------------------------------------------------------------------------------------------------------------- | ------ | ----------------------------------------------------------------------- |
| 8.4.1 | Save Preference (general / situational) | RU | `src/openhuman/agent/tools/save_preference_tests.rs` | ✅ | `save_preference` tool → `user_pref_{general,situational}`, topic-keyed |
| 8.4.2 | Lane A — Standing Prefs in System Prompt | RU | `src/openhuman/learning/prompt_sections.rs`, `src/openhuman/agent/harness/session/turn_tests.rs` | ✅ | General prefs rendered into the system prompt at thread start |
| 8.4.3 | Lane B — Situational Recall (vector-gated) | RU | `src/openhuman/memory/store/unified/query_tests.rs::recall_relevant_by_vector_gates_on_similarity` | ✅ | Per-turn; relevant query injects, unrelated suppresses |
| 8.4.4 | Same-Topic Contradiction (replace) | RU | `src/openhuman/agent/tools/save_preference_tests.rs::recategorising_moves_pref_between_namespaces` | ✅ | `ON CONFLICT REPLACE`; a topic lives in exactly one scope |
| 8.4.5 | Cross-Topic Contradiction Surfacing | RU | `src/openhuman/agent/tools/save_preference_tests.rs::save_surfaces_related_preference_for_contradiction_check` | ✅ | Related prefs surfaced in the tool result for the chat agent to resolve |
| 8.4.6 | vector_chunks Model-Signature Recall Guard | RU | `src/openhuman/memory/store/unified/query_tests.rs::vector_recall_excludes_other_model_signature` | ✅ | Excludes cross-model vectors; dim-guards legacy rows |
---
@@ -347,30 +355,30 @@ Canonical mapping of every product feature to its test source(s). Drives gap-fil
### 10.1 Integration Setup
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ------------------- | ----- | ---------------------------------------------------- | ------ | ------ |
| 10.1.1 | Telegram Connection | WD | `telegram-flow.spec.ts` | ✅ | |
| 10.1.2 | WhatsApp Connection | WD | `app/test/e2e/specs/whatsapp-flow.spec.ts` | ✅ | Was ❌ |
| 10.1.3 | Gmail Connection | WD | `gmail-flow.spec.ts` | ✅ | |
| 10.1.4 | Slack Connection | WD | `app/test/e2e/specs/slack-flow.spec.ts` | ✅ | Was ❌ |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ------------------- | ----- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| 10.1.1 | Telegram Connection | WD | `telegram-flow.spec.ts` | ✅ | |
| 10.1.2 | WhatsApp Connection | WD | `app/test/e2e/specs/whatsapp-flow.spec.ts` | ✅ | Was ❌ |
| 10.1.3 | Gmail Connection | WD | `gmail-flow.spec.ts` | ✅ | |
| 10.1.4 | Slack Connection | WD | `app/test/e2e/specs/slack-flow.spec.ts` | ✅ | Was ❌ |
| 10.1.5 | Yuanbao Connection | RU | `src/openhuman/channels/providers/yuanbao/`, `src/openhuman/channels/controllers/ops.rs::tests::connect_yuanbao_*`, `src/openhuman/channels/runtime/startup.rs::yuanbao_secret_tests` | 🟡 | New API-key channel for Tencent Yuanbao. RU covers sign-token preflight (valid/invalid creds, env-override cluster routing), credentials store hydration (incl. stale app_key guard), and WS reconnect/shutdown. No WDIO spec yet — connect-flow UI is rendered via the generic `ChannelSetupModal` already exercised by other channel flow specs. |
### 10.2 Authentication & Authorization
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ------------------------------------- | ----- | --------------------------------------------------------- | ------ | --------------------------------- |
| 10.2.1 | OAuth / API Token Handling | WD | `skill-oauth.spec.ts` | ✅ | |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ------------------------------------- | ----- | ----------------------------------------------------------------------------- | ------ | ------------------------------------------------------------------------- |
| 10.2.1 | OAuth / API Token Handling | WD | `skill-oauth.spec.ts` | ✅ | |
| 10.2.2 | Scope Selection (Read/Write/Initiate) | WD | `gmail-flow.spec.ts`, `skill-oauth.spec.ts`, `composio-triggers-flow.spec.ts` | 🟡 | Multi-scope matrix not exhaustive; Gmail trigger OAuth read scope covered |
| 10.2.3 | Token Storage & Encryption | RU | `src/openhuman/encryption/`, `src/openhuman/credentials/` | ✅ | |
| 10.2.3 | Token Storage & Encryption | RU | `src/openhuman/encryption/`, `src/openhuman/credentials/` | ✅ | |
### 10.3 Message Sync & Ingestion
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ------------------------- | ----- | ----------------------------------------------------- | ------ | ----- |
| 10.3.1 | Incoming Message Sync | RU+WD | `src/openhuman/channels/tests/`, `gmail-flow.spec.ts` | ✅ | |
| 10.3.2 | Message Deduplication | RU | `src/openhuman/channels/tests/` | ✅ | |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ------------------------- | ----- | ------------------------------------------------------------------------------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------------------------------------- |
| 10.3.1 | Incoming Message Sync | RU+WD | `src/openhuman/channels/tests/`, `gmail-flow.spec.ts` | ✅ | |
| 10.3.2 | Message Deduplication | RU | `src/openhuman/channels/tests/` | ✅ | |
| 10.3.3 | WhatsApp Agent Retrieval | RU | `src/openhuman/whatsapp_data/tools/`, `tests/json_rpc_e2e.rs::whatsapp_data_agent_tools_e2e_1341` | ✅ | Three read-only agent tools wrap the local SQLite store; ingest stays internal-only. See [`docs/whatsapp-data-flow.md`](whatsapp-data-flow.md). |
| 10.3.4 | Real-Time vs Delayed Sync | RU | `src/openhuman/channels/tests/runtime_dispatch.rs` | ✅ | |
| 10.3.4 | Real-Time vs Delayed Sync | RU | `src/openhuman/channels/tests/runtime_dispatch.rs` | ✅ | |
### 10.4 Messaging Operations
@@ -383,11 +391,11 @@ Canonical mapping of every product feature to its test source(s). Drives gap-fil
### 10.5 Cross-Channel Behavior
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ---------------------- | ----- | ------------------------------------------ | ------ | -------------------- |
| 10.5.1 | Channel Isolation | RU | `src/openhuman/channels/tests/identity.rs` | ✅ | |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ---------------------- | ----- | --------------------------------------------------------- | ------ | ----------------------------------------------------------------------------------- |
| 10.5.1 | Channel Isolation | RU | `src/openhuman/channels/tests/identity.rs` | ✅ | |
| 10.5.2 | Unified Inbox Handling | WD+RI | `channels-smoke.spec.ts`, `tests/worker_c_modules_e2e.rs` | 🟡 | UI assertion shallow; RI covers config-only channel status after connect/disconnect |
| 10.5.3 | Context Preservation | RU | `src/openhuman/channels/tests/context.rs` | ✅ | |
| 10.5.3 | Context Preservation | RU | `src/openhuman/channels/tests/context.rs` | ✅ | |
### 10.6 Permission Enforcement
@@ -399,12 +407,12 @@ Canonical mapping of every product feature to its test source(s). Drives gap-fil
### 10.7 Disconnect & Re-Setup
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ---------------------- | ----- | ------------------------------------------- | ------ | -------------------------------- |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ---------------------- | ----- | ----------------------------------------------------- | ------ | ------------------------------------------------------------------- |
| 10.7.1 | Integration Disconnect | WD+RI | `gmail-flow.spec.ts`, `tests/worker_c_modules_e2e.rs` | ✅ | RI covers `channels_disconnect` clearing config-only iMessage state |
| 10.7.2 | Token Revocation | RU | `src/openhuman/credentials/` | ✅ | |
| 10.7.3 | Re-Authorization Flow | WD | `skill-oauth.spec.ts` | 🟡 | Re-auth post-revoke not asserted |
| 10.7.4 | Permission Re-Sync | WD | _missing_ — tracked #968 | ❌ | |
| 10.7.2 | Token Revocation | RU | `src/openhuman/credentials/` | ✅ | |
| 10.7.3 | Re-Authorization Flow | WD | `skill-oauth.spec.ts` | 🟡 | Re-auth post-revoke not asserted |
| 10.7.4 | Permission Re-Sync | WD | _missing_ — tracked #968 | ❌ | |
---
@@ -412,26 +420,26 @@ Canonical mapping of every product feature to its test source(s). Drives gap-fil
### 11.1 Analysis Engine
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | -------------------------- | ----- | ------------------------------------------------------------------------------------------------------------------- | ------ | ----------------------------------------------------------------------------------------- |
| 11.1.1 | Multi-Source Analysis | RI | `tests/memory_graph_sync_e2e.rs` | 🟡 | Frontend trigger untested |
| 11.1.2 | Actionable Item Extraction | VU | `app/src/components/intelligence/__tests__/utils.test.ts` | ✅ | Was ❌ |
| 11.1.3 | Analyze Trigger | WD | `app/test/e2e/specs/insights-dashboard.spec.ts` mounts the route; explicit analyze-handler invocation TBD | 🟡 | Route mounts and search/filter UI assert — full analyze trigger flow tracked as follow-up |
| 11.1.4 | MCP server (stdio + HTTP) | RU | `src/openhuman/mcp_server/` | ✅ | Stdio framing plus Streamable HTTP/SSE session lifecycle; `McpHttpClient` round-trip tests |
| 11.1.5 | Global tool registry | RI | `src/openhuman/tool_registry/`, `tests/json_rpc_e2e.rs`, `tests/domain_modules_e2e.rs`, `tests/worker_b_domain_e2e.rs` | ✅ | Read-only MCP/controller discovery with routes, schemas, version, allowed agents, and health |
| 11.1.6 | SearXNG MCP search | RU | `src/openhuman/integrations/searxng.rs`, `src/openhuman/mcp_server/tools.rs`, `src/openhuman/tools/schemas.rs` | ✅ | Self-hosted search config, normalized results, MCP argument validation, and mocked HTTP execution |
| 11.1.7 | Bundled prompt resources | RU | `src/openhuman/mcp_server/resources.rs`, `src/openhuman/mcp_server/protocol.rs` | ✅ | `resources/list` catalog + `resources/read` happy path, -32002 unknown URI, -32602 missing param, catalog-mirrors-BUILTINS parity test |
| 11.1.8 | Resource templates list | RU | `src/openhuman/mcp_server/resources.rs`, `src/openhuman/mcp_server/protocol.rs` | ✅ | `resources/templates/list` returns `{resourceTemplates: []}` (static catalog), tolerates unknown/cursor params |
| 11.1.10 | MCP registry install→connect→tool_call | RI | `tests/json_rpc_e2e.rs` (`mcp_clients_install_connect_tool_call_happy_path`), `tests/mcp_registry_e2e.rs`, `src/openhuman/mcp_registry/setup_ops.rs` (#3039) | ✅ | HTTP-RPC happy path install→connect→tool_call→update_env against `test-mcp-stub`; transport-aware install (stdio + http_remote) via `build_install_transport` |
| 11.1.11 | MCP env reconfigure + registry creds | RI/VU | `tests/json_rpc_e2e.rs` (`mcp_clients_registry_settings_roundtrip`), `src/openhuman/mcp_registry/registries/mcp_official.rs`, `app/src/components/channels/mcp/InstalledServerDetail.test.tsx` (#3039) | ✅ | `update_env` persist+reconnect; `registry_settings` get/set with secrets write-only (config-first, env-fallback); reconfigure form validation |
| 11.1.12 | MCP UI surface + setup-agent client | VU | `app/src/components/channels/mcp/InstallDialog.test.tsx`, `app/src/services/api/mcpClientsApi.test.ts`, `app/src/services/api/mcpSetupApi.test.ts` (#3039) | ✅ | Skills `?tab=mcp` renders `McpServersTab` (not Coming Soon); auto-connect on install (best-effort); typed `mcpSetupApi` wrapper |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------- | -------------------------------------- | ----- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| 11.1.1 | Multi-Source Analysis | RI | `tests/memory_graph_sync_e2e.rs` | 🟡 | Frontend trigger untested |
| 11.1.2 | Actionable Item Extraction | VU | `app/src/components/intelligence/__tests__/utils.test.ts` | ✅ | Was ❌ |
| 11.1.3 | Analyze Trigger | WD | `app/test/e2e/specs/insights-dashboard.spec.ts` mounts the route; explicit analyze-handler invocation TBD | 🟡 | Route mounts and search/filter UI assert — full analyze trigger flow tracked as follow-up |
| 11.1.4 | MCP server (stdio + HTTP) | RU | `src/openhuman/mcp_server/` | ✅ | Stdio framing plus Streamable HTTP/SSE session lifecycle; `McpHttpClient` round-trip tests |
| 11.1.5 | Global tool registry | RI | `src/openhuman/tool_registry/`, `tests/json_rpc_e2e.rs`, `tests/domain_modules_e2e.rs`, `tests/worker_b_domain_e2e.rs` | ✅ | Read-only MCP/controller discovery with routes, schemas, version, allowed agents, and health |
| 11.1.6 | SearXNG MCP search | RU | `src/openhuman/integrations/searxng.rs`, `src/openhuman/mcp_server/tools.rs`, `src/openhuman/tools/schemas.rs` | ✅ | Self-hosted search config, normalized results, MCP argument validation, and mocked HTTP execution |
| 11.1.7 | Bundled prompt resources | RU | `src/openhuman/mcp_server/resources.rs`, `src/openhuman/mcp_server/protocol.rs` | ✅ | `resources/list` catalog + `resources/read` happy path, -32002 unknown URI, -32602 missing param, catalog-mirrors-BUILTINS parity test |
| 11.1.8 | Resource templates list | RU | `src/openhuman/mcp_server/resources.rs`, `src/openhuman/mcp_server/protocol.rs` | ✅ | `resources/templates/list` returns `{resourceTemplates: []}` (static catalog), tolerates unknown/cursor params |
| 11.1.10 | MCP registry install→connect→tool_call | RI | `tests/json_rpc_e2e.rs` (`mcp_clients_install_connect_tool_call_happy_path`), `tests/mcp_registry_e2e.rs`, `src/openhuman/mcp_registry/setup_ops.rs` (#3039) | ✅ | HTTP-RPC happy path install→connect→tool_call→update_env against `test-mcp-stub`; transport-aware install (stdio + http_remote) via `build_install_transport` |
| 11.1.11 | MCP env reconfigure + registry creds | RI/VU | `tests/json_rpc_e2e.rs` (`mcp_clients_registry_settings_roundtrip`), `src/openhuman/mcp_registry/registries/mcp_official.rs`, `app/src/components/channels/mcp/InstalledServerDetail.test.tsx` (#3039) | ✅ | `update_env` persist+reconnect; `registry_settings` get/set with secrets write-only (config-first, env-fallback); reconfigure form validation |
| 11.1.12 | MCP UI surface + setup-agent client | VU | `app/src/components/channels/mcp/InstallDialog.test.tsx`, `app/src/services/api/mcpClientsApi.test.ts`, `app/src/services/api/mcpSetupApi.test.ts` (#3039) | ✅ | Skills `?tab=mcp` renders `McpServersTab` (not Coming Soon); auto-connect on install (best-effort); typed `mcpSetupApi` wrapper |
<!-- 11.1.9 Vault Markdown Writes — removed: Knowledge Vaults dropped (vault domain + VaultPanel deleted). -->
### 11.2 Insights Dashboard
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ------------------ | ----- | -------------------------------------- | ------ | ------ |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ------------------ | ----- | ---------------------------- | ------ | ------ |
| 11.2.1 | Memory View | WD | `insights-dashboard.spec.ts` | ✅ | Was ❌ |
| 11.2.2 | Source Filtering | WD | `insights-dashboard.spec.ts` | ✅ | Was ❌ |
| 11.2.3 | Search & Retrieval | WD | `insights-dashboard.spec.ts` | ✅ | Was ❌ |
@@ -448,19 +456,19 @@ Canonical mapping of every product feature to its test source(s). Drives gap-fil
### 12.1 Role Unlocking
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ------------------------ | ----- | --------------------------------------------------------------------------------------------------------------------- | ------ | -------------------------------------------------------------------- |
| 12.1.1 | Activity-Based Unlock | VU+WD | `app/src/store/__tests__/rewardsSlice.test.ts`, `app/test/e2e/specs/rewards-unlock-flow.spec.ts` | ✅ | Was ❌ — streak/feature-driven unlock branch |
| 12.1.2 | Integration-Based Unlock | VU+WD | same | ✅ | Was ❌ — Discord membership → role assignment branch |
| 12.1.3 | Plan-Based Unlock | VU+WD | same | ✅ | Was ❌ — plan tier + active subscription branch |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ------------------------ | ----- | ------------------------------------------------------------------------------------------------ | ------ | ---------------------------------------------------- |
| 12.1.1 | Activity-Based Unlock | VU+WD | `app/src/store/__tests__/rewardsSlice.test.ts`, `app/test/e2e/specs/rewards-unlock-flow.spec.ts` | ✅ | Was ❌ — streak/feature-driven unlock branch |
| 12.1.2 | Integration-Based Unlock | VU+WD | same | ✅ | Was ❌ — Discord membership → role assignment branch |
| 12.1.3 | Plan-Based Unlock | VU+WD | same | ✅ | Was ❌ — plan tier + active subscription branch |
### 12.2 Progress Tracking
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ---------------------- | ----- | ----------------------------------------------------------------------------------------------------------------------------- | ------ | ---------------------------------------------------------------------------------------------------- |
| 12.2.1 | Message Count Tracking | VU+WD | `rewardsSlice.test.ts`, `rewards-progression-persistence.spec.ts` | ✅ | Was ❌ — message-driven progress proxied by `metrics.featuresUsedCount` (no literal field) |
| 12.2.2 | Usage Metrics | VU+WD | same | ✅ | Was ❌ — current streak + cumulative tokens |
| 12.2.3 | State Persistence | VU+WD | same | ✅ | Was ❌ — restart-equivalent (page unmount + remount + re-fetch); admin request log asserts re-fetch |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ---------------------- | ----- | ----------------------------------------------------------------- | ------ | --------------------------------------------------------------------------------------------------- |
| 12.2.1 | Message Count Tracking | VU+WD | `rewardsSlice.test.ts`, `rewards-progression-persistence.spec.ts` | ✅ | Was ❌ — message-driven progress proxied by `metrics.featuresUsedCount` (no literal field) |
| 12.2.2 | Usage Metrics | VU+WD | same | ✅ | Was ❌ — current streak + cumulative tokens |
| 12.2.3 | State Persistence | VU+WD | same | ✅ | Was ❌ — restart-equivalent (page unmount + remount + re-fetch); admin request log asserts re-fetch |
---
@@ -468,45 +476,45 @@ Canonical mapping of every product feature to its test source(s). Drives gap-fil
### 13.1 Account & Security
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ------------------ | ----- | -------------------------------------------------------------------- | ------ | --------------------- |
| 13.1.1 | Profile Management | VU | `app/src/components/settings/panels/__tests__/PrivacyPanel.test.tsx` | 🟡 | |
| 13.1.2 | Linked Accounts | WD | `auth-access-control.spec.ts` | 🟡 | UI surface unasserted |
| 13.1.3 | Meet Handoff Prompt-Injection Guard | VU | `app/src/services/__tests__/webviewAccountService.meetPromptInjection.test.ts` | ✅ | Was ❌ — guard blocks handoff on hostile transcripts and wraps non-blocked transcripts in `<meeting_transcript source="untrusted_external_audio">` delimiters (#1920) |
| 13.1.4 | Wallet Balances Panel | VU | `app/src/components/settings/panels/__tests__/WalletBalancesPanel.test.tsx`, `app/src/services/walletApi.test.ts` | ✅ | Loading/error/empty/loaded states; Retry + Refresh re-invocation; chain badges; truncated address; providerStatus chip |
| 13.1.5 | Approval History | VU | `app/src/components/settings/panels/__tests__/ApprovalHistoryPanel.test.tsx`, `app/src/services/api/approvalApi.test.ts` | ✅ | Was ❌ — read-only audit surface over `approval_list_recent_decisions`; covers loaded/empty/error/refresh states, per-decision badge, and the bare-array vs `{result,logs}` envelope normalization |
| 13.1.6 | Action Timeout | VU, RU, RI | `app/src/components/settings/panels/__tests__/AgentAccessPanel.test.tsx`, `src/openhuman/tool_timeout/mod.rs`, `src/openhuman/config/ops_tests.rs`, `tests/json_rpc_e2e.rs` | ✅ | New (#3100) — UI control over `config_get/update_agent_settings` for `[agent].agent_timeout_secs`; covers load/persist-on-blur/range-rejection/no-op/env-override-disable, the runtime-mutable `tool_timeout` resolver + env precedence, ops apply/reject, and the RPC roundtrip |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ----------------------------------- | ---------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| 13.1.1 | Profile Management | VU | `app/src/components/settings/panels/__tests__/PrivacyPanel.test.tsx` | 🟡 | |
| 13.1.2 | Linked Accounts | WD | `auth-access-control.spec.ts` | 🟡 | UI surface unasserted |
| 13.1.3 | Meet Handoff Prompt-Injection Guard | VU | `app/src/services/__tests__/webviewAccountService.meetPromptInjection.test.ts` | ✅ | Was ❌ — guard blocks handoff on hostile transcripts and wraps non-blocked transcripts in `<meeting_transcript source="untrusted_external_audio">` delimiters (#1920) |
| 13.1.4 | Wallet Balances Panel | VU | `app/src/components/settings/panels/__tests__/WalletBalancesPanel.test.tsx`, `app/src/services/walletApi.test.ts` | ✅ | Loading/error/empty/loaded states; Retry + Refresh re-invocation; chain badges; truncated address; providerStatus chip |
| 13.1.5 | Approval History | VU | `app/src/components/settings/panels/__tests__/ApprovalHistoryPanel.test.tsx`, `app/src/services/api/approvalApi.test.ts` | ✅ | Was ❌ — read-only audit surface over `approval_list_recent_decisions`; covers loaded/empty/error/refresh states, per-decision badge, and the bare-array vs `{result,logs}` envelope normalization |
| 13.1.6 | Action Timeout | VU, RU, RI | `app/src/components/settings/panels/__tests__/AgentAccessPanel.test.tsx`, `src/openhuman/tool_timeout/mod.rs`, `src/openhuman/config/ops_tests.rs`, `tests/json_rpc_e2e.rs` | ✅ | New (#3100) — UI control over `config_get/update_agent_settings` for `[agent].agent_timeout_secs`; covers load/persist-on-blur/range-rejection/no-op/env-override-disable, the runtime-mutable `tool_timeout` resolver + env precedence, ops apply/reject, and the RPC roundtrip |
### 13.2 Automation & Channels
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | --------------------- | ----- | ----------------------------------------------------------- | ------ | ----- |
| 13.2.1 | Channel Configuration | WD | `app/test/e2e/specs/settings-channels-permissions.spec.ts` | ✅ | |
| 13.2.2 | Permission Settings | WD | `app/test/e2e/specs/settings-channels-permissions.spec.ts` | ✅ | |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | --------------------- | ----- | ---------------------------------------------------------- | ------ | ----- |
| 13.2.1 | Channel Configuration | WD | `app/test/e2e/specs/settings-channels-permissions.spec.ts` | ✅ | |
| 13.2.2 | Permission Settings | WD | `app/test/e2e/specs/settings-channels-permissions.spec.ts` | ✅ | |
### 13.3 AI & Skills
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ------------------- | ----- | ------------------------------------------------------------------------------------------------------------------------- | ------ | ----------------------------------- |
| 13.3.1 | Model Configuration | VU+WD | `app/src/components/settings/panels/__tests__/AutocompletePanel.test.tsx`, `app/test/e2e/specs/settings-ai-skills.spec.ts` | ✅ | AI-model-switch covered |
| 13.3.2 | Skill Toggle | WD | `skill-lifecycle.spec.ts`, `app/test/e2e/specs/settings-ai-skills.spec.ts` | ✅ | |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ------------------- | ----- | -------------------------------------------------------------------------------------------------------------------------- | ------ | ----------------------- |
| 13.3.1 | Model Configuration | VU+WD | `app/src/components/settings/panels/__tests__/AutocompletePanel.test.tsx`, `app/test/e2e/specs/settings-ai-skills.spec.ts` | ✅ | AI-model-switch covered |
| 13.3.2 | Skill Toggle | WD | `skill-lifecycle.spec.ts`, `app/test/e2e/specs/settings-ai-skills.spec.ts` | ✅ | |
### 13.4 Developer Options
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ------------------ | ----- | ---------------------------------------------------- | ------ | ----- |
| 13.4.1 | Webhook Inspection | WD | `app/test/e2e/specs/settings-dev-options.spec.ts` | ✅ | |
| 13.4.2 | Runtime Logs | WD | `app/test/e2e/specs/settings-dev-options.spec.ts` | ✅ | |
| 13.4.3 | Memory Debug | WD | `app/test/e2e/specs/settings-dev-options.spec.ts` | ✅ | |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ------------------ | ----- | ------------------------------------------------- | ------ | ----- |
| 13.4.1 | Webhook Inspection | WD | `app/test/e2e/specs/settings-dev-options.spec.ts` | ✅ | |
| 13.4.2 | Runtime Logs | WD | `app/test/e2e/specs/settings-dev-options.spec.ts` | ✅ | |
| 13.4.3 | Memory Debug | WD | `app/test/e2e/specs/settings-dev-options.spec.ts` | ✅ | |
### 13.5 Data Management
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ---------------- | ----- | ------------------------------------------------------- | ------ | -------------------------------------- |
| 13.5.1 | Clear App Data | WD | `app/test/e2e/specs/settings-data-management.spec.ts` | ✅ | Destructive — confirm-then-reset |
| 13.5.2 | Cache Reset | WD | `app/test/e2e/specs/settings-data-management.spec.ts` | ✅ | |
| 13.5.3 | Full State Reset | WD | `app/test/e2e/specs/settings-data-management.spec.ts` | ✅ | Restart-and-verify fresh-install state |
| 13.5.4 | Migration from another assistant (OpenClaw) | VU+RU | `app/src/components/settings/panels/__tests__/MigrationPanel.test.tsx`, `src/openhuman/migration/ops.rs` (existing) | ✅ | Was ❌ — UI now wraps the existing `openhuman.migrate_openclaw` RPC with preview-then-apply + confirm. Hermes tracked as follow-up under #1440 |
| ID | Feature | Layer | Test path(s) | Status | Notes |
| ------ | ------------------------------------------- | ----- | ------------------------------------------------------------------------------------------------------------------- | ------ | ---------------------------------------------------------------------------------------------------------------------------------------------- |
| 13.5.1 | Clear App Data | WD | `app/test/e2e/specs/settings-data-management.spec.ts` | ✅ | Destructive — confirm-then-reset |
| 13.5.2 | Cache Reset | WD | `app/test/e2e/specs/settings-data-management.spec.ts` | ✅ | |
| 13.5.3 | Full State Reset | WD | `app/test/e2e/specs/settings-data-management.spec.ts` | ✅ | Restart-and-verify fresh-install state |
| 13.5.4 | Migration from another assistant (OpenClaw) | VU+RU | `app/src/components/settings/panels/__tests__/MigrationPanel.test.tsx`, `src/openhuman/migration/ops.rs` (existing) | ✅ | Was ❌ — UI now wraps the existing `openhuman.migrate_openclaw` RPC with preview-then-apply + confirm. Hermes tracked as follow-up under #1440 |
---
src/openhuman/agent/harness/subagent_runner/ops/loop_.rs
+5 -1
View File
@@ -60,6 +60,10 @@ pub(super) async fn run_inner_loop(
handoff_cache: Option<&ResultHandoffCache>,
parent: &ParentExecutionContext,
extended_policy: bool,
// Optional steering channel. When `Some`, the child engine drains
// steer/collect messages at iteration boundaries so the parent can
// `steer_subagent` a running async sub-agent. `None` = non-steerable.
run_queue: Option<std::sync::Arc<crate::openhuman::agent::harness::run_queue::RunQueue>>,
) -> Result<(String, usize, AggregatedUsage, Option<String>), SubagentRunError> {
// An autonomous skill run (set via `with_autonomous_iter_cap`) lifts the
// per-agent cap so sub-agents run until done / the circuit breaker trips.
@@ -199,7 +203,7 @@ pub(super) async fn run_inner_loop(
max_iterations,
None, // sub-agents don't stream a draft
&["ask_user_clarification"],
None, // sub-agents don't support run-queue steering
run_queue, // steering channel for `steer_subagent` (None = non-steerable)
)),
)
.await?;
src/openhuman/agent/harness/subagent_runner/ops/runner.rs
+1
View File
@@ -696,6 +696,7 @@ async fn run_typed_mode(
handoff_cache.as_deref(),
parent,
definition.iteration_policy == IterationPolicy::Extended,
options.run_queue.clone(),
))
.await?;
src/openhuman/agent/harness/subagent_runner/ops_tests.rs
+63
View File
@@ -179,6 +179,7 @@ fn append_subagent_role_contract_is_idempotent() {
// ── End-to-end runner tests with mock provider ────────────────────────
use crate::openhuman::agent::harness::fork_context::with_parent_context;
use crate::openhuman::agent::harness::run_queue::{QueueMode, QueuedMessage, RunQueue};
use crate::openhuman::inference::provider::{
ChatRequest as PChatRequest, ChatResponse, Provider, ProviderDelta, ToolCall,
};
@@ -480,6 +481,7 @@ async fn typed_mode_returns_text_through_runner() {
initial_history: None,
checkpoint_dir: None,
worktree_action_dir: None,
run_queue: None,
},
)
.await
@@ -493,6 +495,66 @@ async fn typed_mode_returns_text_through_runner() {
assert_eq!(outcome.task_id, "t1");
}
#[tokio::test]
async fn run_queue_steer_lands_in_subagent_history() {
// End-to-end proof that flipping the subagent loop's run-queue arg from
// `None` to `Some(queue)` wires steering all the way through: a message
// pushed to the queue before the run is drained by the inner
// `run_turn_engine` and appears as a `[User steering message]:` user turn
// in the exact request sent to the provider. This is the mechanism behind
// the `steer_subagent` tool.
let provider = ScriptedProvider::new(vec![text_response("acknowledged")]);
let parent = make_parent(provider.clone(), vec![stub("file_read")]);
let def = make_def_named_tools(&[]);
let run_queue = RunQueue::new();
run_queue
.push(QueuedMessage {
text: "switch focus to memory safety".into(),
mode: QueueMode::Steer,
client_id: "steer_subagent".into(),
thread_id: "t-steer".into(),
queued_at_ms: 0,
model_override: None,
temperature: None,
profile_id: None,
locale: None,
})
.await;
let outcome = with_parent_context(parent, async {
run_subagent(
&def,
"investigate the bug",
SubagentRunOptions {
task_id: Some("t-steer".into()),
run_queue: Some(run_queue),
..Default::default()
},
)
.await
})
.await
.expect("runner should succeed");
assert_eq!(outcome.output, "acknowledged");
let captured = provider.captured.lock();
let steered = captured[0]
.messages
.iter()
.any(|m| m.role == "user" && m.content.contains("switch focus to memory safety"));
assert!(
steered,
"steer message should be injected into the sub-agent's first request, got: {:?}",
captured[0]
.messages
.iter()
.map(|m| (&m.role, &m.content))
.collect::<Vec<_>>()
);
}
#[tokio::test]
async fn typed_mode_no_memory_context_in_user_message() {
// Verifies that sub-agents skip memory loading entirely: the
@@ -592,6 +654,7 @@ async fn typed_mode_filters_tools_by_skill_filter() {
initial_history: None,
checkpoint_dir: None,
worktree_action_dir: None,
run_queue: None,
},
)
.await
src/openhuman/agent/harness/subagent_runner/types.rs
+7
View File
@@ -66,6 +66,13 @@ pub struct SubagentRunOptions {
/// When `None` (the default), behaviour is unchanged — tools fall through
/// to `security.action_dir`.
pub worktree_action_dir: Option<PathBuf>,
/// Steering channel for a running (typically async) sub-agent. When set,
/// the inner `run_turn_engine` drains steer/collect messages from this
/// queue at iteration boundaries — exactly like the main agent loop — so
/// the parent can `steer_subagent` mid-flight. `None` keeps today's
/// non-steerable behaviour.
pub run_queue: Option<std::sync::Arc<crate::openhuman::agent::harness::run_queue::RunQueue>>,
}
/// Terminal status of a sub-agent run.
src/openhuman/agent/tools/delegate_to_personality.rs
+1
View File
@@ -218,6 +218,7 @@ impl Tool for DelegateToPersonalityTool {
initial_history: None,
checkpoint_dir: None,
worktree_action_dir: None,
run_queue: None,
};
match run_subagent(&definition, &prompt, options).await {
src/openhuman/agent_orchestration/mod.rs
+1
View File
@@ -8,6 +8,7 @@
pub mod agent_teams;
pub mod command_center;
mod ops;
pub mod running_subagents;
pub mod tools;
pub mod types;
pub mod workflow_runs;
src/openhuman/agent_orchestration/ops.rs
+1
View File
@@ -465,6 +465,7 @@ impl AgentOrchestrationSession {
initial_history: None,
checkpoint_dir: None,
worktree_action_dir: None,
run_queue: None,
};
let task_session = self.clone();
src/openhuman/agent_orchestration/running_subagents.rs
+423
View File
@@ -0,0 +1,423 @@
//! Registry of in-flight async sub-agents that can be **steered** mid-run.
//!
//! `spawn_async_subagent` runs a child as a detached `tokio` task. On its own
//! that task is opaque: the parent gets a `task_id` back but has no channel into
//! the running loop and no way to collect the result inline. This registry
//! closes both gaps.
//!
//! Each running async sub-agent registers, keyed by its `task_id`, with:
//! - an `Arc<RunQueue>` — the same steering channel the inner `run_turn_engine`
//! drains at iteration boundaries, so `steer_subagent` can inject a message;
//! - a `watch::Receiver<SubagentStatus>` — so `wait_subagent` can block until the
//! child reaches a terminal status;
//! - an `AbortHandle` — kept for a future `close_agent` tool.
//!
//! Ownership is enforced: only the spawning parent (matched by `parent_session`)
//! may steer or wait on a given sub-agent. Terminal entries are pruned on `wait`,
//! and swept on `register` only once the table passes a soft cap, so it can't
//! grow unbounded if a parent never waits (the Codex "spawn-slot leak" failure
//! mode — openai/codex#18335).
use std::collections::HashMap;
use std::sync::{Arc, Mutex, OnceLock};
use std::time::Duration;
use tokio::sync::watch;
use tokio::task::AbortHandle;
use crate::openhuman::agent::harness::run_queue::{QueueMode, QueuedMessage, RunQueue};
/// Terminal/transient state of a running async sub-agent, published by the
/// spawner's background task and observed by `wait_subagent`.
#[derive(Debug, Clone)]
pub enum SubagentStatus {
/// Still executing its inner tool-call loop.
Running,
/// Finished normally with a final response.
Completed { output: String, iterations: usize },
/// Paused on `ask_user_clarification`; resume via `continue_subagent`.
AwaitingUser { question: String },
/// The run errored out.
Failed { error: String },
}
impl SubagentStatus {
pub fn is_terminal(&self) -> bool {
!matches!(self, SubagentStatus::Running)
}
}
struct RunningSubagentEntry {
#[allow(dead_code)]
agent_id: String,
parent_session: String,
run_queue: Arc<RunQueue>,
abort: AbortHandle,
status: watch::Receiver<SubagentStatus>,
}
/// Soft cap on registry size. Terminal entries are only swept when the table
/// grows past this, so the common case (a handful of live sub-agents) never
/// evicts a still-uncollected terminal result out from under a `wait`/`steer`.
const REGISTRY_SOFT_CAP: usize = 256;
static REGISTRY: OnceLock<Mutex<HashMap<String, RunningSubagentEntry>>> = OnceLock::new();
fn registry() -> &'static Mutex<HashMap<String, RunningSubagentEntry>> {
REGISTRY.get_or_init(|| Mutex::new(HashMap::new()))
}
/// Create the status channel a spawner threads into [`register`].
///
/// The spawner moves the [`watch::Sender`] into its detached task and `send`s a
/// terminal [`SubagentStatus`] on completion. Dropping the sender (e.g. a
/// panicked/aborted task) closes the channel, which `wait_subagent` surfaces as
/// a failure rather than hanging.
pub fn status_channel() -> (
watch::Sender<SubagentStatus>,
watch::Receiver<SubagentStatus>,
) {
watch::channel(SubagentStatus::Running)
}
/// Register a running async sub-agent so it can be steered and waited on.
///
/// Call this *after* `tokio::spawn` so the [`AbortHandle`] is available; the
/// task owns the matching [`watch::Sender`] from [`status_channel`]. Once the
/// table passes [`REGISTRY_SOFT_CAP`], registration sweeps already-terminal
/// entries so it stays bounded even if a parent never calls `wait_subagent`.
pub fn register(
task_id: String,
agent_id: String,
parent_session: String,
run_queue: Arc<RunQueue>,
abort: AbortHandle,
status: watch::Receiver<SubagentStatus>,
) {
let entry = RunningSubagentEntry {
agent_id,
parent_session,
run_queue,
abort,
status,
};
let mut map = registry().lock().expect("running_subagents mutex poisoned");
if map.len() >= REGISTRY_SOFT_CAP {
// Only under genuine pressure: sweep collected/terminal entries so the
// table can't grow without bound when a parent never waits (the Codex
// spawn-slot leak). Live (Running) entries are always retained.
map.retain(|_, e| !e.status.borrow().is_terminal());
}
map.insert(task_id.clone(), entry);
log::debug!(
"[running_subagents] registered task_id={} live_entries={}",
task_id,
map.len()
);
}
/// Why a steer could not be delivered.
#[derive(Debug, PartialEq, Eq)]
pub enum SteerError {
/// No such sub-agent — never existed, or already finished and pruned.
Unknown,
/// The caller's `parent_session` does not own this sub-agent.
NotOwned,
/// The sub-agent already reached a terminal status.
AlreadyDone,
}
/// Inject a message into a running sub-agent's steering queue. The child's
/// `run_turn_engine` drains it at the next iteration boundary.
pub async fn steer(
task_id: &str,
parent_session: &str,
text: String,
mode: QueueMode,
) -> Result<(), SteerError> {
let run_queue = {
let map = registry().lock().expect("running_subagents mutex poisoned");
let entry = map.get(task_id).ok_or(SteerError::Unknown)?;
if entry.parent_session != parent_session {
return Err(SteerError::NotOwned);
}
if entry.status.borrow().is_terminal() {
return Err(SteerError::AlreadyDone);
}
entry.run_queue.clone()
};
run_queue
.push(QueuedMessage {
text,
mode,
client_id: "steer_subagent".to_string(),
thread_id: task_id.to_string(),
queued_at_ms: now_ms(),
model_override: None,
temperature: None,
profile_id: None,
locale: None,
})
.await;
log::info!(
"[running_subagents] steered task_id={} mode={}",
task_id,
mode
);
Ok(())
}
/// Why a wait could not be set up.
#[derive(Debug, PartialEq, Eq)]
pub enum WaitError {
Unknown,
NotOwned,
}
/// Result of waiting on a sub-agent.
#[derive(Debug)]
pub enum WaitOutcome {
/// The sub-agent reached a terminal status (entry pruned).
Terminal(SubagentStatus),
/// The timeout elapsed first; the entry is left intact so the parent can
/// wait again. Carries the latest (non-terminal) status snapshot.
TimedOut(SubagentStatus),
}
/// Block until `task_id` reaches a terminal status or `timeout` elapses.
pub async fn wait(
task_id: &str,
parent_session: &str,
timeout: Duration,
) -> Result<WaitOutcome, WaitError> {
let mut rx = {
let map = registry().lock().expect("running_subagents mutex poisoned");
let entry = map.get(task_id).ok_or(WaitError::Unknown)?;
if entry.parent_session != parent_session {
return Err(WaitError::NotOwned);
}
entry.status.clone()
};
// Fast path: already terminal.
let current = rx.borrow_and_update().clone();
if current.is_terminal() {
prune(task_id);
return Ok(WaitOutcome::Terminal(current));
}
let waited = async {
loop {
if rx.changed().await.is_err() {
// Sender dropped without a terminal status (task aborted/panicked).
return SubagentStatus::Failed {
error: "sub-agent task ended without reporting a result".to_string(),
};
}
let status = rx.borrow().clone();
if status.is_terminal() {
return status;
}
}
};
match tokio::time::timeout(timeout, waited).await {
Ok(status) => {
prune(task_id);
Ok(WaitOutcome::Terminal(status))
}
Err(_) => Ok(WaitOutcome::TimedOut(rx.borrow().clone())),
}
}
/// Abort a running sub-agent and drop its registry entry. Kept for a future
/// `close_agent` tool; the abort handle is stored at spawn time.
pub fn close(task_id: &str, parent_session: &str) -> bool {
let mut map = registry().lock().expect("running_subagents mutex poisoned");
match map.get(task_id) {
Some(entry) if entry.parent_session == parent_session => {
entry.abort.abort();
map.remove(task_id);
true
}
_ => false,
}
}
fn prune(task_id: &str) {
registry()
.lock()
.expect("running_subagents mutex poisoned")
.remove(task_id);
}
fn now_ms() -> u64 {
std::time::SystemTime::now()
.duration_since(std::time::UNIX_EPOCH)
.map(|d| d.as_millis() as u64)
.unwrap_or(0)
}
#[cfg(test)]
mod tests {
use super::*;
fn dummy_abort() -> AbortHandle {
tokio::spawn(async {}).abort_handle()
}
/// Register a sub-agent for tests, returning the status sender so the test
/// can drive completion. Keeping the sender alive keeps the channel open.
fn register_test(
task_id: &str,
parent_session: &str,
rq: Arc<RunQueue>,
) -> watch::Sender<SubagentStatus> {
let (tx, rx) = status_channel();
register(
task_id.into(),
"researcher".into(),
parent_session.into(),
rq,
dummy_abort(),
rx,
);
tx
}
#[tokio::test]
async fn steer_pushes_into_the_subagent_queue() {
let rq = RunQueue::new();
let tx = register_test("task-steer", "session-A", rq.clone());
steer(
"task-steer",
"session-A",
"refocus on memory safety".into(),
QueueMode::Steer,
)
.await
.expect("steer should succeed");
let status = rq.status().await;
assert_eq!(status.steers, 1, "steer should land in the steer lane");
// collect mode goes to the collect lane
steer(
"task-steer",
"session-A",
"extra context".into(),
QueueMode::Collect,
)
.await
.unwrap();
assert_eq!(rq.status().await.collects, 1);
let _ = tx.send(SubagentStatus::Completed {
output: "done".into(),
iterations: 1,
});
prune("task-steer");
}
#[tokio::test]
async fn steer_rejects_cross_parent_and_unknown() {
let rq = RunQueue::new();
let _tx = register_test("task-owned", "session-owner", rq);
assert_eq!(
steer(
"task-owned",
"session-intruder",
"x".into(),
QueueMode::Steer
)
.await,
Err(SteerError::NotOwned)
);
assert_eq!(
steer(
"task-missing",
"session-owner",
"x".into(),
QueueMode::Steer
)
.await,
Err(SteerError::Unknown)
);
prune("task-owned");
}
#[tokio::test]
async fn steer_after_terminal_is_rejected() {
let rq = RunQueue::new();
let tx = register_test("task-term", "session-A", rq);
let _ = tx.send(SubagentStatus::Failed {
error: "boom".into(),
});
assert_eq!(
steer("task-term", "session-A", "x".into(), QueueMode::Steer).await,
Err(SteerError::AlreadyDone)
);
prune("task-term");
}
#[tokio::test]
async fn wait_returns_completion_once_published() {
let rq = RunQueue::new();
let tx = register_test("task-wait", "session-A", rq);
tokio::spawn(async move {
let _ = tx.send(SubagentStatus::Completed {
output: "the answer".into(),
iterations: 3,
});
// keep sender alive until after send
drop(tx);
});
let outcome = wait("task-wait", "session-A", Duration::from_secs(5))
.await
.expect("wait should resolve");
match outcome {
WaitOutcome::Terminal(SubagentStatus::Completed { output, iterations }) => {
assert_eq!(output, "the answer");
assert_eq!(iterations, 3);
}
other => panic!("expected completed terminal, got {other:?}"),
}
// pruned after a terminal wait
assert!(matches!(
wait("task-wait", "session-A", Duration::from_millis(10)).await,
Err(WaitError::Unknown)
));
}
#[tokio::test]
async fn wait_times_out_and_leaves_entry_intact() {
let rq = RunQueue::new();
let _tx = register_test("task-slow", "session-A", rq);
let outcome = wait("task-slow", "session-A", Duration::from_millis(20))
.await
.expect("wait should resolve");
assert!(matches!(
outcome,
WaitOutcome::TimedOut(SubagentStatus::Running)
));
// still steerable after a timed-out wait
assert!(steer(
"task-slow",
"session-A",
"still here".into(),
QueueMode::Steer
)
.await
.is_ok());
prune("task-slow");
}
}
src/openhuman/agent_orchestration/tools.rs
+6
View File
@@ -14,9 +14,13 @@ mod spawn_parallel_agents;
mod spawn_subagent;
#[path = "tools/spawn_worker_thread.rs"]
pub mod spawn_worker_thread;
#[path = "tools/steer_subagent.rs"]
mod steer_subagent;
#[cfg(test)]
#[path = "tools/tools_e2e_tests.rs"]
mod tools_e2e_tests;
#[path = "tools/wait_subagent.rs"]
mod wait_subagent;
#[path = "tools/worker_thread.rs"]
mod worker_thread;
@@ -29,3 +33,5 @@ pub use spawn_async_subagent::SpawnAsyncSubagentTool;
pub use spawn_parallel_agents::SpawnParallelAgentsTool;
pub use spawn_subagent::SpawnSubagentTool;
pub use spawn_worker_thread::SpawnWorkerThreadTool;
pub use steer_subagent::SteerSubagentTool;
pub use wait_subagent::WaitSubagentTool;
src/openhuman/agent_orchestration/tools/continue_subagent.rs
+1
View File
@@ -229,6 +229,7 @@ impl Tool for ContinueSubagentTool {
initial_history: Some(history),
checkpoint_dir: Some(checkpoint_dir.clone()),
worktree_action_dir: None,
run_queue: None,
};
// Run the sub-agent from its checkpoint
src/openhuman/agent_orchestration/tools/dispatch.rs
+1
View File
@@ -109,6 +109,7 @@ pub(crate) async fn dispatch_subagent(
initial_history: None,
checkpoint_dir: None,
worktree_action_dir: None,
run_queue: None,
};
match run_subagent(definition, prompt, options).await {
src/openhuman/agent_orchestration/tools/spawn_async_subagent.rs
+38 -2
View File
@@ -7,10 +7,12 @@
use crate::core::event_bus::{publish_global, DomainEvent};
use crate::openhuman::agent::harness::definition::AgentDefinitionRegistry;
use crate::openhuman::agent::harness::fork_context::{current_parent, with_parent_context};
use crate::openhuman::agent::harness::run_queue::RunQueue;
use crate::openhuman::agent::harness::subagent_runner::{
run_subagent, SubagentRunOptions, SubagentRunStatus,
};
use crate::openhuman::agent::progress::AgentProgress;
use crate::openhuman::agent_orchestration::running_subagents::{self, SubagentStatus};
use crate::openhuman::tools::traits::{PermissionLevel, Tool, ToolResult};
use async_trait::async_trait;
use serde_json::json;
@@ -225,6 +227,14 @@ impl Tool for SpawnAsyncSubagentTool {
.await;
}
// Steering channel + status channel so the parent can `steer_subagent`
// this run mid-flight and `wait_subagent` for its result. The engine
// drains `steer_queue` at iteration boundaries; `status_tx` publishes
// the terminal state to any waiter.
let steer_queue = RunQueue::new();
let task_queue = steer_queue.clone();
let (status_tx, status_rx) = running_subagents::status_channel();
let background_parent = parent.clone();
let background_definition = definition.clone();
let background_agent_id = definition.id.clone();
@@ -234,7 +244,7 @@ impl Tool for SpawnAsyncSubagentTool {
let background_worker_thread_id = worker_thread_id.clone();
let background_prompt = add_background_contract(&prompt);
tokio::spawn(async move {
let join = tokio::spawn(async move {
let options = SubagentRunOptions {
skill_filter_override: None,
toolkit_override,
@@ -245,6 +255,7 @@ impl Tool for SpawnAsyncSubagentTool {
initial_history: None,
checkpoint_dir: None,
worktree_action_dir: None,
run_queue: Some(task_queue),
};
let result = with_parent_context(background_parent, async move {
@@ -255,6 +266,11 @@ impl Tool for SpawnAsyncSubagentTool {
match result {
Ok(outcome) => match outcome.status {
SubagentRunStatus::Completed => {
// Unblock `wait_subagent` with the final output first.
let _ = status_tx.send(SubagentStatus::Completed {
output: outcome.output.clone(),
iterations: outcome.iterations,
});
publish_global(DomainEvent::SubagentCompleted {
parent_session: background_parent_session,
task_id: outcome.task_id.clone(),
@@ -276,6 +292,9 @@ impl Tool for SpawnAsyncSubagentTool {
}
}
SubagentRunStatus::AwaitingUser { question, .. } => {
let _ = status_tx.send(SubagentStatus::AwaitingUser {
question: question.clone(),
});
let error = format!(
"async sub-agent requested user clarification and was not continued: {question}"
);
@@ -298,6 +317,9 @@ impl Tool for SpawnAsyncSubagentTool {
},
Err(err) => {
let error = err.to_string();
let _ = status_tx.send(SubagentStatus::Failed {
error: error.clone(),
});
publish_global(DomainEvent::SubagentFailed {
parent_session: background_parent_session,
task_id: background_task_id.clone(),
@@ -317,6 +339,17 @@ impl Tool for SpawnAsyncSubagentTool {
}
});
// Register *after* spawn so the AbortHandle is available. The task owns
// `status_tx`; this side holds `status_rx` for `wait_subagent`.
running_subagents::register(
task_id.clone(),
definition.id.clone(),
parent_session.clone(),
steer_queue,
join.abort_handle(),
status_rx,
);
let payload = json!({
"task_id": task_id,
"agent_id": definition.id,
@@ -324,8 +357,11 @@ impl Tool for SpawnAsyncSubagentTool {
"worker_thread_id": worker_thread_id,
});
Ok(ToolResult::success(format!(
"Accepted background sub-agent `{}`. Do not wait for it before answering the user.\n\n[async_subagent_ref]\n{}\n[/async_subagent_ref]",
"Accepted background sub-agent `{}` (task_id `{}`). Do not block on it before answering the user. \
You may redirect it mid-run with `steer_subagent {{ task_id, message }}` and collect its result \
with `wait_subagent {{ task_id }}`.\n\n[async_subagent_ref]\n{}\n[/async_subagent_ref]",
payload["agent_id"].as_str().unwrap_or("subagent"),
task_id,
serde_json::to_string(&payload).unwrap_or_else(|_| "{}".to_string())
)))
}
src/openhuman/agent_orchestration/tools/spawn_parallel_agents.rs
+1
View File
@@ -682,6 +682,7 @@ async fn run_one_parallel_task(
initial_history: None,
checkpoint_dir: None,
worktree_action_dir: worktree_path.clone(),
run_queue: None,
};
let run_result = run_subagent(&definition, &prompt, options).await;
src/openhuman/agent_orchestration/tools/spawn_subagent.rs
+1
View File
@@ -459,6 +459,7 @@ impl Tool for SpawnSubagentTool {
initial_history: None,
checkpoint_dir: None,
worktree_action_dir: None,
run_queue: None,
};
let progress_sink = current_parent().and_then(|p| p.on_progress.clone());
src/openhuman/agent_orchestration/tools/spawn_worker_thread.rs
+1
View File
@@ -228,6 +228,7 @@ impl Tool for SpawnWorkerThreadTool {
initial_history: None,
checkpoint_dir: None,
worktree_action_dir: None,
run_queue: None,
};
tracing::debug!(
src/openhuman/agent_orchestration/tools/steer_subagent.rs
+174
View File
@@ -0,0 +1,174 @@
//! Tool: `steer_subagent` — inject a message into a running async sub-agent.
//!
//! Pairs with `spawn_async_subagent`: that tool returns a `task_id` for a child
//! running in the background. `steer_subagent` pushes a message into that child's
//! steering queue, which its inner `run_turn_engine` drains at the next iteration
//! boundary — so the parent can redirect or feed data to a running sub-agent
//! without waiting for it to finish or restarting it. Mirrors Codex `send_input`.
use crate::openhuman::agent::harness::fork_context::current_parent;
use crate::openhuman::agent::harness::run_queue::QueueMode;
use crate::openhuman::agent_orchestration::running_subagents::{self, SteerError};
use crate::openhuman::tools::traits::{PermissionLevel, Tool, ToolResult};
use async_trait::async_trait;
use serde_json::json;
pub struct SteerSubagentTool;
impl SteerSubagentTool {
pub fn new() -> Self {
Self
}
}
impl Default for SteerSubagentTool {
fn default() -> Self {
Self::new()
}
}
#[async_trait]
impl Tool for SteerSubagentTool {
fn name(&self) -> &str {
"steer_subagent"
}
fn description(&self) -> &str {
"Send a message into a running async sub-agent (one you started with \
spawn_async_subagent), redirecting or feeding it data mid-run without \
restarting it. The sub-agent picks the message up at its next step. Use \
`mode: steer` (default) for a new instruction it must address, or \
`mode: collect` for silent extra context. Returns immediately; use \
wait_subagent to collect the final result."
}
fn parameters_schema(&self) -> serde_json::Value {
json!({
"type": "object",
"required": ["task_id", "message"],
"properties": {
"task_id": {
"type": "string",
"description": "The task_id returned by spawn_async_subagent (see its [async_subagent_ref] block)."
},
"message": {
"type": "string",
"description": "Instruction or data to inject into the running sub-agent."
},
"mode": {
"type": "string",
"enum": ["steer", "collect"],
"default": "steer",
"description": "steer = a new instruction the sub-agent must address; collect = silent additional context."
}
}
})
}
fn permission_level(&self) -> PermissionLevel {
PermissionLevel::Execute
}
async fn execute(&self, args: serde_json::Value) -> anyhow::Result<ToolResult> {
let task_id = args
.get("task_id")
.and_then(|v| v.as_str())
.unwrap_or("")
.trim()
.to_string();
let message = args
.get("message")
.and_then(|v| v.as_str())
.unwrap_or("")
.trim()
.to_string();
let mode = match args.get("mode").and_then(|v| v.as_str()).unwrap_or("steer") {
"collect" => QueueMode::Collect,
_ => QueueMode::Steer,
};
if task_id.is_empty() {
return Ok(ToolResult::error("steer_subagent: `task_id` is required"));
}
if message.is_empty() {
return Ok(ToolResult::error("steer_subagent: `message` is required"));
}
let parent_session = match current_parent() {
Some(parent) => parent.session_id,
None => {
return Ok(ToolResult::error(
"steer_subagent called outside of an agent turn",
));
}
};
log::info!(
"[steer_subagent] task_id={} mode={} chars={}",
task_id,
mode,
message.chars().count()
);
match running_subagents::steer(&task_id, &parent_session, message, mode).await {
Ok(()) => Ok(ToolResult::success(format!(
"Steered sub-agent `{task_id}` ({mode}). It will pick this up at its next step. \
Use wait_subagent {{ task_id: \"{task_id}\" }} to collect its result."
))),
Err(SteerError::Unknown) => Ok(ToolResult::error(format!(
"steer_subagent: no running sub-agent with task_id `{task_id}`. It may have already \
finished — use wait_subagent to collect its result, or check the task_id."
))),
Err(SteerError::NotOwned) => Ok(ToolResult::error(format!(
"steer_subagent: sub-agent `{task_id}` was not started by this agent and cannot be steered."
))),
Err(SteerError::AlreadyDone) => Ok(ToolResult::error(format!(
"steer_subagent: sub-agent `{task_id}` has already finished. Use wait_subagent to collect its result."
))),
}
}
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn schema_requires_task_id_and_message() {
let schema = SteerSubagentTool::new().parameters_schema();
let required = schema
.get("required")
.and_then(|v| v.as_array())
.expect("required list");
assert!(required.iter().any(|v| v.as_str() == Some("task_id")));
assert!(required.iter().any(|v| v.as_str() == Some("message")));
}
#[tokio::test]
async fn missing_task_id_is_rejected() {
let tool = SteerSubagentTool::new();
let res = tool.execute(json!({ "message": "go" })).await.unwrap();
assert!(res.is_error);
assert!(res.output().contains("task_id"));
}
#[tokio::test]
async fn missing_message_is_rejected() {
let tool = SteerSubagentTool::new();
let res = tool.execute(json!({ "task_id": "sub-1" })).await.unwrap();
assert!(res.is_error);
assert!(res.output().contains("message"));
}
#[tokio::test]
async fn outside_agent_turn_is_rejected() {
// No PARENT_CONTEXT task-local installed in a bare test.
let tool = SteerSubagentTool::new();
let res = tool
.execute(json!({ "task_id": "sub-1", "message": "go" }))
.await
.unwrap();
assert!(res.is_error);
assert!(res.output().contains("outside of an agent turn"));
}
}
src/openhuman/agent_orchestration/tools/wait_subagent.rs
+171
View File
@@ -0,0 +1,171 @@
//! Tool: `wait_subagent` — block until a running async sub-agent finishes.
//!
//! Pairs with `spawn_async_subagent` / `steer_subagent`: once the parent has
//! fanned out background work, `wait_subagent` collects a child's final result
//! inline (with a timeout), instead of relying solely on lifecycle events.
//! Mirrors Codex `wait`.
use std::time::Duration;
use crate::openhuman::agent::harness::fork_context::current_parent;
use crate::openhuman::agent_orchestration::running_subagents::{
self, SubagentStatus, WaitError, WaitOutcome,
};
use crate::openhuman::tools::traits::{PermissionLevel, Tool, ToolResult};
use async_trait::async_trait;
use serde_json::json;
const DEFAULT_TIMEOUT_SECS: u64 = 120;
const MAX_TIMEOUT_SECS: u64 = 600;
pub struct WaitSubagentTool;
impl WaitSubagentTool {
pub fn new() -> Self {
Self
}
}
impl Default for WaitSubagentTool {
fn default() -> Self {
Self::new()
}
}
#[async_trait]
impl Tool for WaitSubagentTool {
fn name(&self) -> &str {
"wait_subagent"
}
fn description(&self) -> &str {
"Block until an async sub-agent (started with spawn_async_subagent) \
finishes, then return its final result. Optionally bound the wait with \
`timeout_secs` (default 120, max 600); on timeout it reports the \
sub-agent is still running and you can call wait_subagent again."
}
fn parameters_schema(&self) -> serde_json::Value {
json!({
"type": "object",
"required": ["task_id"],
"properties": {
"task_id": {
"type": "string",
"description": "The task_id returned by spawn_async_subagent."
},
"timeout_secs": {
"type": "integer",
"minimum": 1,
"maximum": MAX_TIMEOUT_SECS,
"description": "Max seconds to block before returning a 'still running' result. Default 120."
}
}
})
}
fn permission_level(&self) -> PermissionLevel {
PermissionLevel::Execute
}
async fn execute(&self, args: serde_json::Value) -> anyhow::Result<ToolResult> {
let task_id = args
.get("task_id")
.and_then(|v| v.as_str())
.unwrap_or("")
.trim()
.to_string();
if task_id.is_empty() {
return Ok(ToolResult::error("wait_subagent: `task_id` is required"));
}
let timeout_secs = args
.get("timeout_secs")
.and_then(|v| v.as_u64())
.unwrap_or(DEFAULT_TIMEOUT_SECS)
.clamp(1, MAX_TIMEOUT_SECS);
let parent_session = match current_parent() {
Some(parent) => parent.session_id,
None => {
return Ok(ToolResult::error(
"wait_subagent called outside of an agent turn",
));
}
};
log::info!(
"[wait_subagent] task_id={} timeout_secs={}",
task_id,
timeout_secs
);
match running_subagents::wait(
&task_id,
&parent_session,
Duration::from_secs(timeout_secs),
)
.await
{
Ok(WaitOutcome::Terminal(SubagentStatus::Completed { output, iterations })) => {
Ok(ToolResult::success(format!(
"Sub-agent `{task_id}` completed in {iterations} iteration(s):\n\n{output}"
)))
}
Ok(WaitOutcome::Terminal(SubagentStatus::AwaitingUser { question })) => {
Ok(ToolResult::success(format!(
"Sub-agent `{task_id}` paused for clarification and did not finish: {question}\n\n\
It cannot proceed unattended. Resume it with continue_subagent once you have an answer."
)))
}
Ok(WaitOutcome::Terminal(SubagentStatus::Failed { error })) => Ok(ToolResult::error(
format!("Sub-agent `{task_id}` failed: {error}"),
)),
// `Running` is never terminal; treat defensively as a timeout-style result.
Ok(WaitOutcome::Terminal(SubagentStatus::Running))
| Ok(WaitOutcome::TimedOut(_)) => Ok(ToolResult::success(format!(
"Sub-agent `{task_id}` is still running after {timeout_secs}s. \
Continue with other work and call wait_subagent again later, or steer_subagent to redirect it."
))),
Err(WaitError::Unknown) => Ok(ToolResult::error(format!(
"wait_subagent: no sub-agent with task_id `{task_id}`. It may have already finished and \
been collected, or the task_id is wrong."
))),
Err(WaitError::NotOwned) => Ok(ToolResult::error(format!(
"wait_subagent: sub-agent `{task_id}` was not started by this agent."
))),
}
}
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn schema_requires_task_id() {
let schema = WaitSubagentTool::new().parameters_schema();
let required = schema
.get("required")
.and_then(|v| v.as_array())
.expect("required list");
assert!(required.iter().any(|v| v.as_str() == Some("task_id")));
}
#[tokio::test]
async fn missing_task_id_is_rejected() {
let res = WaitSubagentTool::new().execute(json!({})).await.unwrap();
assert!(res.is_error);
assert!(res.output().contains("task_id"));
}
#[tokio::test]
async fn outside_agent_turn_is_rejected() {
let res = WaitSubagentTool::new()
.execute(json!({ "task_id": "sub-1" }))
.await
.unwrap();
assert!(res.is_error);
assert!(res.output().contains("outside of an agent turn"));
}
}
src/openhuman/tools/ops.rs
+7
View File
@@ -159,6 +159,13 @@ pub fn all_tools_with_runtime(
// `agent::harness::subagent_runner` for the dispatch path.
Box::new(SpawnSubagentTool::new()),
Box::new(SpawnAsyncSubagentTool::new()),
// Steer a running async sub-agent mid-flight and collect its result:
// `steer_subagent { task_id, message }` injects into the child's
// run-queue (drained at its next iteration boundary), `wait_subagent
// { task_id }` blocks for the final output. See
// `agent_orchestration::running_subagents`.
Box::new(SteerSubagentTool::new()),
Box::new(WaitSubagentTool::new()),
Box::new(ContinueSubagentTool::new()),
Box::new(SpawnParallelAgentsTool::new()),
Box::new(DelegateToPersonalityTool::new()),
tests/inference_agent_raw_coverage_e2e.rs
+1
View File
@@ -4642,6 +4642,7 @@ async fn agent_subagent_public_types_cover_task_local_and_error_display_paths()
initial_history: None,
checkpoint_dir: None,
worktree_action_dir: None,
run_queue: None,
};
assert_eq!(options.skill_filter_override.as_deref(), Some("docs"));
assert_eq!(options.toolkit_override.as_deref(), Some("github"));