The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes repository code diffs and can use external LLM providers to analyze them, creating a surface where instructions embedded in the code could potentially influence the agent's behavior.\n
Ingestion points: The skill reads local repository content using git diff as described in SKILL.md.\n
Boundary markers: The instructions do not specify delimiters or warnings for the agent to ignore instructions found within the processed code.\n
Capability inventory: The skill executes shell commands via npx, git, and gh across its workflow.\n
Sanitization: There is no mention of filtering or sanitizing the diff content before analysis.\n- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to download and execute the @coldtea/pr-lens-cli package from the npm registry using npx at runtime.\n
Evidence: Found in SKILL.md as npx @coldtea/pr-lens-cli@latest.\n- [DATA_EXFILTRATION]: The skill includes an optional feature to upload diagram data to a third-party visualization service for sharing.\n
Evidence: The canvas push command in SKILL.md uploads generated graph data to prlens.dev.

pr-lens