feat: agent eval mode for loop exit conditions#126

Merged

dean0x merged 17 commits intomainfrom

feat-agent-eval-mode

Mar 30, 2026

Commits on Mar 30, 2026

Loop loop-06fd8b6e-0085-4e90-9085-8f42c3f28568 iteration 1 — pass
Dean Sharon
committed
style: fix biome formatting and import organization
Dean Sharon
committed
fix: address self-review issues
Dean Sharon
committed
fix: address shepherd review gaps
Dean Sharon
committed
refactor: simplify agent evaluator and cleanup
Dean Sharon
committed
refactor(loop-handler): extract cleanupIterationTracking helper and scope stale guard to agent mode

Dean Sharon
and
claude
committed
fix(schedule-manager): allow agent-mode scheduled loops with empty exitCondition

Dean Sharon
and
claude
committed
fix(loops): cancel orphaned eval task when parent loop is cancelled

Dean Sharon
and
claude
committed
docs: document agent eval mode and fix EvalMode type consistency

Dean Sharon
and
claude
committed
refactor(loops): complete EvalMode enum rollout and add exhaustive switch

Dean Sharon
and
claude
committed
style(mcp-adapter): restore default EvalMode.SHELL on Zod schemas

Dean Sharon
and
claude
committed
fix(loop-manager): add runtime guard for invalid evalMode values

Dean Sharon
and
claude
committed
style: fix biome formatting after parallel resolver changes
Dean Sharon
committed
refactor(loop): extract IterationResultFields type and reduce test boilerplate

Dean Sharon
and
claude
committed
refactor: simplify cleanup and remove redundant schedule update
Dean Sharon
committed
fix: resolve pre-existing review issues #3, #4, #5
Dean Sharon
committed
fix: resolve Greptile PR review findings (P1 CLI bug, P2 eval prompt)
Dean Sharon
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: agent eval mode for loop exit conditions#126

feat: agent eval mode for loop exit conditions#126
dean0x merged 17 commits intomainfrom
feat-agent-eval-mode

Commits on Mar 30, 2026

Loop loop-06fd8b6e-0085-4e90-9085-8f42c3f28568 iteration 1 — pass

style: fix biome formatting and import organization

fix: address self-review issues

fix: address shepherd review gaps

refactor: simplify agent evaluator and cleanup

refactor(loop-handler): extract cleanupIterationTracking helper and scope stale guard to agent mode

fix(schedule-manager): allow agent-mode scheduled loops with empty exitCondition

fix(loops): cancel orphaned eval task when parent loop is cancelled

docs: document agent eval mode and fix EvalMode type consistency

refactor(loops): complete EvalMode enum rollout and add exhaustive switch

style(mcp-adapter): restore default EvalMode.SHELL on Zod schemas

fix(loop-manager): add runtime guard for invalid evalMode values

style: fix biome formatting after parallel resolver changes

refactor(loop): extract IterationResultFields type and reduce test boilerplate

refactor: simplify cleanup and remove redundant schedule update

fix: resolve pre-existing review issues #3, #4, #5

fix: resolve Greptile PR review findings (P1 CLI bug, P2 eval prompt)