The Agent Skills Directory

[SAFE]: The skill acts as a documentation-heavy reference for secure coding practices and does not contain any executable code or risky commands within its own structure.
[INDIRECT_PROMPT_INJECTION]: The skill explicitly acknowledges the risk of processing untrusted repository text and tool outputs. It instructs the agent to treat such data as untrusted evidence and to enforce security policies outside of model-generated prose.
[EXTERNAL_DOWNLOADS]: The skill references documentation and repositories from well-known organizations including OpenAI, Cloudflare, the W3C, and the Model Context Protocol (MCP) for provenance and further reading. These references are documented neutrally as sources of best practices.

security-implementation