The Agent Skills Directory

[SAFE]: The skill is composed of purely instructional markdown text. It does not include any scripts, executable binaries, or system configuration changes.
[PROMPT_INJECTION]: The skill requires the agent to analyze external proposals and repository evidence, which creates a surface for indirect prompt injection. Ingestion points: User-provided plans, specs, and repository files (code, tests, docs). Boundary markers: Absent; no delimiters or safety warnings are included for processed content. Capability inventory: File reading for evidence gathering. Sanitization: Absent; no escaping or validation of ingested content is specified.

grill-me