The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were detected in the skill's logic or resource requests.
[INDIRECT_PROMPT_INJECTION]: The skill creates an attack surface by ingesting data from external logs and plans without sanitization. Ingestion points: Reads files from production/playtests/ and production/qa/qa-plan-*.md. Boundary markers: None identified in the prompt template. Capability inventory: Uses the Write tool to create Markdown documentation. Sanitization: None; data is interpolated directly into the protocol template.

soak-test