dexter/BTerminal
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

docs: update meta files for reviewer agent role

Browse source
This commit is contained in:
Hibryda 2026-03-12 00:54:43 +01:00
parent 323bb1b040
commit d1a4d9f220
4 changed files with 5 additions and 7 deletions
Download patch file Download diff file Expand all files Collapse all files

5
TODO.md
View file

@ -8,11 +8,9 @@
- [ ] **Multi-machine TLS/certificate pinning** -- TLS support for bterminal-relay + certificate pinning in RemoteManager.
- [ ] **Agent Teams real-world testing** -- Env var whitelist fix done. 3 test sessions ran ($1.10, $0.69, $1.70) but model didn't spawn subagents — needs complex multi-part prompts to trigger delegation. Test with CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1.
### Dexter's Roadmap (Tribunal S-3 Hybrid)
- [ ] **Reviewer agent role** -- Tier 1 specialist with role='reviewer'. #review-queue btmsg channel on task→review transition. Verdict as bttask comment + #review-log channel. review_queue_depth in attention scoring (10pts/stale review, cap 50). GroupAgentRole already includes 'reviewer'.
## Completed
- [x] **Reviewer agent role** -- Tier 1 specialist with role='reviewer'. Reviewer workflow in agent-prompts.ts (8-step process). #review-queue/#review-log auto-channels in bttask.rs (auto-post on task→review). reviewQueueDepth in attention scoring (10pts/task, cap 50). Tasks tab for reviewer in ProjectBox (reuses TaskBoardTab). 10s review queue polling. 388 vitest + 76 cargo. | Done: 2026-03-12
- [x] **Auto-wake Manager** -- wake-scheduler.svelte.ts + wake-scorer.ts (24 tests). 3 user-selectable strategies: persistent (resume prompt), on-demand (fresh session), smart (threshold-gated). 6 signals: AttentionSpike(1.0), ContextPressureCluster(0.9), BurnRateAnomaly(0.8), TaskQueuePressure(0.7), ReviewBacklog(0.6), PeriodicFloor(0.1). Settings UI: strategy segmented button + threshold slider. GroupAgentConfig: wakeStrategy + wakeThreshold fields. 381 vitest + 72 cargo. | Done: 2026-03-12
- [x] **Dashboard metrics panel** -- MetricsPanel.svelte: new ProjectBox tab ('metrics', PERSISTED-LAZY, all projects). Live view: fleet aggregates, project health grid, task board summary (5 kanban columns, 10s poll), attention queue. History view: 5 SVG sparkline charts (cost/tokens/turns/tools/duration), stats row, session table. 25 tests. 357 vitest + 72 cargo. | Done: 2026-03-12
- [x] **Brand Dexter's new types (SOLID Phase 3b)** -- Added GroupId + AgentId branded types to ids.ts. Applied to ~40 sites: groups.ts interfaces, btmsg-bridge.ts (5 interfaces, 15 function params), bttask-bridge.ts (2 interfaces, 6 params), groups-bridge.ts (3 interfaces), GroupAgentsPanel, TaskBoardTab, SettingsTab. agentToProject() uses `as unknown as ProjectId` cast. 3 test files updated. 332 vitest + 72 cargo. | Done: 2026-03-11
@ -22,4 +20,3 @@
- [x] **SOLID Phase 2 — agent-dispatcher.ts split** -- 496→260 lines. Extracted 4 modules: utils/worktree-detection.ts (pure function, 5 tests), utils/session-persistence.ts (session maps + persist), utils/auto-anchoring.ts (compaction anchor), utils/subagent-router.ts (spawn + route). Dispatcher is thin coordinator. 286 vitest + 49 cargo tests. | Done: 2026-03-11
- [x] **SOLID Phase 2 — session.rs split** -- 1008→7 sub-modules under session/ directory (mod.rs, sessions.rs, layout.rs, settings.rs, ssh.rs, agents.rs, metrics.rs, anchors.rs). pub(in crate::session) conn visibility. 21 new cargo tests. 49 cargo tests total. | Done: 2026-03-11
- [x] **SOLID Phase 1 Refactoring** -- Extracted AttentionScorer pure function (14 tests), shared str()/num() type guards, split lib.rs (976→170 lines, 11 command modules). 286 vitest + 49 cargo tests. | Done: 2026-03-11
- [x] **Configurable stall threshold** -- Per-project range slider (5–60 min, step 5) in SettingsTab. `stallThresholdMin` in ProjectConfig, `setStallThreshold()` API in health store, ProjectBox $effect sync. Adaptive suggestions deferred (needs 50+ sessions in session_metrics). | Done: 2026-03-11