retrieval-judge

What problem does it solve?

After recall_search, this Skill enforces a critical judgment step to filter results by relevance rather than treating them as authoritative, reducing noise in knowledge retrieval.

Core Features & Use Cases

• Evaluation checklist: Checks title/type match, content relevance, and applicability to the current task.
• Auditable decisions: Logs judgments and keeps a clean, reproducible trail for review.
• Use Case: When you run recall_search in a knowledge-assembly workflow, apply this Skill to discard irrelevant results and keep only those that address the user's query.

Quick Start

After performing recall_search, apply the evaluation steps to each result, log the outcome with flight_recorder_log, and proceed with the kept results.