weichen
b45c6abaa7
Record execution agent history
...
Why: retain which CLI handled each task so routing can improve from real outcomes over time
Evidence: AgentExecHistoryTests and git diff --check
Impact: appends sanitized source-external JSONL records without changing run outcomes
Refs: feature/goal-first-delivery
2026-10-15 07:41:51 +00:00
weichen
f93ef567c7
Isolate execution agents from shared Git refs
...
Why: prevent unrelated sibling worktree delivery from invalidating otherwise safe agent attempts
Evidence: 58 focused tests, configuration validation, governance regression, and diff check
Impact: uses private shared clones and copies only host-owned transport refs back to canonical Git
Refs: U5 runtime acceptance; docs/superpowers/plans/2026-10-15-project-agent-execution.md
2026-10-15 06:19:10 +00:00
weichen
8c1649b232
Stop stalled execution agents on idle timeout
...
Why: prevent silent provider stalls from consuming the full execution timeout
Evidence: 57 focused tests, configuration validation, 401-test governance regression, and diff check
Impact: adds backward-compatible idle_timeout_seconds and fail-stop process supervision
Refs: U5 runtime acceptance; docs/superpowers/plans/2026-10-15-project-agent-execution.md
2026-10-15 05:41:37 +00:00
weichen
8f11d002ae
Add recoverable execution agent controller
...
Why: provide isolated, recoverable per-unit CLI execution and safe host fallback
Evidence: 43 focused tests, configuration validation, governance regression, and diff check
Impact: adds agent-exec validate/run controller, receipts, transports, and Make entrypoint
Refs: U3; docs/superpowers/plans/2026-10-15-project-agent-execution.md
2026-10-15 04:46:25 +00:00
weichen