The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from gtm/proof.json and gtm/offer-stack.json, which represent a potential surface for indirect instructions. However, the risk is mitigated through structured data requirements and local script validation.
Ingestion points: The skill reads from gtm/proof.json and gtm/offer-stack.json as specified in SKILL.md.
Boundary markers: Instructions for handling the JSON data are structured, though no explicit prompt delimiters are used for the file contents.
Capability inventory: The skill has access to Read, Write, and a restricted Bash tool that executes the local scripts/score_proof.py script.
Sanitization: The score_proof.py script implements robust validation, including JSON schema checks, type checking, and semantic validation (e.g., checking for specific metrics, timeframe, and banned empty praise strings).

proof-assets