The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process potentially untrusted data from external marketing tools, scraped web content, and customer emails.
Ingestion points: SKILL.md (Data Sources) and references/connectors.md define the surfaces where untrusted data enters the agent context.
Boundary markers: The skill contains explicit instructions in SKILL.md and references/skill-contract.md to treat external content as evidence and ignore any embedded instructions.
Capability inventory: The skill is primarily instructional; real-world side effects like mutations or persistence require explicit user authorization.
Sanitization: The risk is addressed via logical prompting constraints rather than programmatic filtering.
[SAFE]: The skill follows secure credential management practices by explicitly instructing users to keep API keys out of skills and source control. It also includes instructions for respecting Robots.txt and Terms of Service for its connectors.

reactivation-specialist