Loading...
Loading...
Found 5 Skills
Design LLM-as-Judge evaluators for subjective criteria that code-based checks cannot handle. Use when a failure mode requires interpretation (tone, faithfulness, relevance, completeness). Do NOT use when the failure mode can be checked with code (regex, schema validation, execution tests). Do NOT use when you need to validate or calibrate the judge — use validate-evaluator instead.
Capture noteworthy review violations for the eval framework. Use when validator-run finds review failures — judges violations and saves notable ones to evals/inventory.yml.
Cross-check the agent's own answer with independent reviewers before bringing it to the user. Use when the user says 'second opinion', 'sanity check', 'cross-check', 'am I missing something', 'stress-test', 'devil's advocate', 'run a full review', 'this is important', 'high-stakes', 'help me choose between', 'critique this', or similar. Three review styles: Devil's Advocate (single critique), Boardroom Debate (parallel multi-reviewer validation), Round-Table (multi-round brainstorm for choosing between paths). Not for factual lookups — use WebSearch for those.
Use when running a stage-gate or gate review for an AI product, deciding Go/Kill/Pivot at a gate, judging whether an AI initiative is ready to advance to the next stage, triaging a portfolio of AI ideas, or choosing between a full and a fast-track path. Applies a Stage-Gate process built for AI products — incremental investment, dual value-and-technical validation, alignment/ethics/data checks at every gate — to reach an evidence-based gate decision.
verify when running best-of-N, choosing among multiple attempts, a 1-10 or yes/no judge would collapse close traces, comments claim success without a run, or a live attempt may have stalled.