| name | pagekit-evaluator-pass |
| description | Run the adversarial evaluator pass on a completed PageKit run. Post-evaluation skeptical second read. Use when a run has an evaluation.md but no evaluator-pass.md, or when the orchestrator needs an evaluator pass. Delegates to the pagekit-evaluator-pass subagent. |
PageKit Evaluator Pass
You are invoking the adversarial evaluator pass on a completed run.
When to use
- After the run's
evaluation.md is written.
- Before declaring the run fully-logged (the fully-logged tier requires
evaluator-pass.md).
- When a reviewer asks "is this run as good as the evaluation says?"
Read first
../pagekit/references/run-logging.md — where the evaluator pass fits in the run-logging tieragents/pagekit-evaluator-pass.md — the subagent you will delegate to
Inputs
- path to the run folder (e.g.,
runs/kind-bowl-real/)
Procedure
- Confirm the run has
evaluation.md. If not, the evaluator pass is premature — finish the evaluation first.
- Invoke the
pagekit-evaluator-pass subagent. Pass the run-folder path.
- The subagent runs read-only, produces
runs/<name>/evaluator-pass.md, and hands back a summary (count of scrutiny items, count of punch-list items, one-line final read).
- Relay the result to the caller. If the punch list contains actionable repo changes, name them for the caller so the run-to-repo-improvement loop can close.
Quality gate
A strong evaluator pass:
- names specific quiet claims from the run, not general impressions
- writes an "implication" for each scrutiny item (what this suggests for the method or the repo)
- ends with a concrete punch list of file-level changes
- does not reward defensiveness
If the subagent's output is general rather than specific, ask it to retry with more direct reference to the run's artifacts.
Relationship to pagekit-claim-checker
pagekit-claim-checker reviews the draft line by line at a chosen severity.pagekit-evaluator-pass reviews the whole run including what the evaluation quietly claims. Different target, different voice.
