 Godopu
|
7fde1d2b8a
|
docs(loop): merge multi_agent_workflow.md into SKILL.md, drop duplicate
|
2026-07-16 13:36:12 +09:00 |
|
|
Godopu
|
626f35adfc
|
refactor: harden run_loop.sh verdict parser, add atomic promotion, and revise multi_agent_workflow.md guidelines
|
2026-07-16 12:57:56 +09:00 |
|
|
Godopu
|
dfdbf69187
|
fix(loop): correct self-relative REPO_ROOT depth in run_loop.sh
|
2026-07-16 12:06:37 +09:00 |
|
|
Godopu
|
0f970b97fa
|
fix(loop): address planner and reviewer architecture defects
- Fix B-1: Correct Mermaid sequence diagram syntax (fi -> end) in SKILL.md and PLAN_LOOP.md
- Fix B-2: Force target agent exclusion from active reviewers in run_loop.sh and correct creator session role in registry
- Fix B-3 & B-4: Integrate WAIT_TIMEOUT deadline inside wait_for_job
- Fix M-3: Update CHANGES_DIFF to use dynamic cumulative git diff
- Fix M-2: Resolve verdict string parsing and substring collisions
|
2026-07-16 08:28:57 +09:00 |
|
|
Godopu
|
bacf139447
|
feat(skills): add multi-agent-mux-loop autonomous loop orchestrator
Implement a new autonomous planning-execution-review orchestration loop
supporting:
- Collaborative planning & creator-planner challenge discussions (--plan-talk)
- Self-planning and self-review fallbacks
- Custom reviewer target list and all-reviewer unanimous PASS verdicts
- API-cost runaway safety limits via --max-loop
Integrated inside deploy/install.sh checklist and deploy/gitea-ci.yml.
|
2026-07-16 08:12:24 +09:00 |
|