The Agent Skills Directory

[SAFE]: The skill provides instructional guidance for performing workspace security audits. It uses standard system tools like grep and git to search for patterns related to leaked API keys, tokens, and PII. These operations are conducted locally within the agent's workspace.
[SAFE]: No network exfiltration or unauthorized data access patterns were detected. The skill explicitly states that it does not make network requests or send data externally.
[SAFE]: The 'Security Standing Order' and hardening instructions provided are legitimate defensive measures intended to protect the agent from prompt injection and social engineering attacks.
[EXTERNAL_DOWNLOADS]: The skill contains links to the author's website (theagentledger.com) and GitHub repository (github.com/theagentledger/agent-skills) for additional resources and updates. These references are consistent with the skill's metadata and stated purpose.

security-hardening