What problem does it solve? A generic prompt scorecard cannot detect issues specific to prompts that embed context, issue tool directives, or act as specification documents. This Skill layers IL knowledge-base findings on top of a /prompt-evaluator run so those prompt classes get mechanism-level audit coverage without duplicating the baseline rubric. ## Core Features & Use Cases - Trigger-based IL extensions: Fires G2a/G2b Contract checks for context-embedding prompts, G5 checks for tool-directive prompts, and G1 checks for spec documents, each gated by preconditions. - Integrated single report: Merges the /prompt-evaluator 4-discipline scorecard with IL findings, confidence tags, citations, and a tier trace into one output. - Early-exit discipline: Reports that the baseline scorecard is sufficient when no IL trigger fires, avoiding redundant findings. - Use Case: You have a long CLAUDE.md-style system prompt that pre-loads context and routes work to sub-agents. Run this Skill to get the standard scorecard plus Contract-specific findings on context structuring and tool-directive design that the baseline rubric cannot reach. ## Quick Start Run the assess-prompt skill on my system prompt file and include the prior prompt-evaluator output so I get the integrated IL audit report.
