mass-remediate-pattern
Pass
Audited by Gen Agent Trust Hub on Apr 5, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill includes a rule requiring the agent to address the user as "The Brougham 22". This is a benign stylistic instruction and does not attempt to subvert safety filters or manipulate the agent's core instructions.
- [COMMAND_EXECUTION]: The skill instructs the agent to use standard shell utilities such as
grep,sed, andgitfor locating and modifying source code. These are standard operations for code remediation and are used within a structured process that includes isolated testing. - [REMOTE_CODE_EXECUTION]: The instructions involve running established project test suites like
pytestandnpm testto verify the integrity of changes. This is a standard validation procedure and does not involve the execution of untrusted remote scripts or the download of external dependencies.
Audit Metadata