The Agent Skills Directory

[SAFE]: No executable code, scripts, or shell commands are present in the skill. It serves as a static reference for human analysts to build containment plans.
[SAFE]: The allowed-tools configuration is strictly limited to read-only file system operations (Read, Grep, Glob), preventing unauthorized system modifications or network exfiltration.
[SAFE]: The skill incorporates a robust 'Prompt Injection Safety Notice' in Section 8, explicitly instructing the agent to treat attacker-sourced content as data only and to maintain strict role boundaries.
[SAFE]: All external references point to authoritative and trusted organizations, including NIST, MITRE, CISA, and Microsoft. No suspicious or unverified external sources were detected.

containment