Menu
Score generated prose or product output that cleared deterministic checks against a rubric threshold.
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Basé sur la classification professionnelle SOC
Orchestration skill: chains the full ADLC Build Loop. PRD → Brief → Council → Scaffold → Codegen → LDD → TDD → Council → PR. Use when implementing a new feature end-to-end.
Generate a bounded, schema-valid goal prompt from one validated Build Brief task for zero-context external execution.
Applicability-aware Definition of Done checklist. Core checks apply to every task; overlay checks activate only when the applicability manifest says the surface exists. Binary verification — pass or fail. Triggers at Phase 4 completion and Phase 5 entry.
Score whether each task is specific enough for autonomous, one-shot, production-ready execution.
Reviews code changes for understanding, blast radius, state, secrets, assumptions, and explainability. Produces a blocking comprehension artifact when implications are unclear.
Resolve ambiguous triage outcomes in the middle confidence band.
| name | slop-judge |
| description | Score generated prose or product output that cleared deterministic checks against a rubric threshold. |
| contract_version | 2.0.0 |
| side_effect_profile | read_only |
| activation | {"mode":"judgement","consumes_manifest":true,"model_class":"fast_judge","cost_guard":{"max_tokens_per_call":750,"expected_calls_per_run":3}} |
Use this only after stop-slop regex and pattern checks do not hard-fail.
The judge exists to close the output-side quality loop. It is not a prompt rewriter. It scores the produced output against a concrete rubric and returns a numeric result that a gate can compare to a threshold.
{
"mode": "general | outreach | product_output | agent_output",
"regex_screen": "pass",
"content": "string",
"audience": "internal | external",
"rubric": [
"Concrete criterion such as specificity, actionability, schema validity, policy fit, or non-hallucination"
],
"threshold": 0.7,
"baseline_score": 0.82,
"regression_tolerance": 0.03
}
{
"verdict": "pass | revise",
"score": 0.0,
"threshold": 0.7,
"criterion_scores": [
{
"criterion": "string",
"score": 0.0,
"reason": "string"
}
],
"regression_delta": 0.0,
"rationale": "string",
"signals": ["generic_filler", "passive_evasion", "tautology"],
"new_eval_case_candidate": {
"source": "council_rejection | human_edit | runtime_failure | production_sample | other",
"input": "optional string",
"bad_output": "string",
"expected_quality": "string",
"metric": "rubric_score",
"threshold": 0.7
}
}
pass means score >= threshold and any baseline regression is within regression_tolerance.revise means the output scored below threshold, regressed beyond tolerance, or remains padded, evasive, generic, structurally invalid, or self-referential even though deterministic checks did not catch it.score is the average unless the input names a stricter metric.revise with signals: ["missing_benchmark"].new_eval_case_candidate so feedback-loop can promote the failure into the saved suite after approval.