goal-evaluatorlisted

# Goal Evaluator ## Contract Iron law: deterministic checks beat LLM judgment: run every `verification_command` first and never substitute judge output for a runnable command's exit code. Invoked by `/flow:goal evaluate <id>` and `/flow:debug` step 6 (`trigger=command`), and by `hooks/scripts/flow-goal-evaluator.sh` in evaluator-loop mode (`trigger=stop-hook`), with goal id, run id, and trigger. Returns `{verdict, confidence, delta, reason, next_step_hint, criterion_results}` plus, for terminal outcomes, a `proposed_transition`; writes evidence sidecars and non-terminal lifecycle updates, never `last-verdict.json` and never a terminal status. Permitted skips: the judge when no fuzzy criteria remain; the path-boundary check when `constraints.allowed_paths` is unset. ## Inputs 1. **Goal id**: `.flow/goals/<id>.goal.yaml` with status `active` (or resumable `waiting_for_user`, `waiting_for_ci`, `blocked`). 2. **Run id**: for `.flow/runs/<run-id>/evidence/`; defaults to the goal's `scope.run_id`. 3. **Trigger**: `manual | stop-hook | command`. ## Outputs 1. Updated goal: AC `status` (`pending → evidence_collected → pass | fail`), `evidence_ref`, `last_evaluated_at`, `last_result`; `lifecycle.last_evaluation`; non-terminal `lifecycle.status`. 2. `*.evidence.yaml` sidecars via `bin/flow-record-evidence.sh` (the `goal-evidence-ledger` skill). 3. A `goal-evaluation` journal artifact. 4. The structured verdict, returned to the caller. ## Workflow ### Step 1: Load Read the goal