The Agent Skills Directory

[SAFE]: The skill's instructions are purely focused on ensuring that new code is accompanied by ownership metadata, runbooks, and CODEOWNERS entries. No malicious patterns were identified.
[NO_CODE]: The skill does not include any external scripts, binaries, or configuration files, relying entirely on natural language instructions for the agent.
[DATA_EXPOSURE]: There is no evidence of sensitive data access or exfiltration. The skill references standard organizational metadata such as team names and documentation stubs.
[PROMPT_INJECTION]: No attempts to bypass safety filters or override agent behavior were detected.

ownership-chain