surge-recon
Pass
Audited by Gen Agent Trust Hub on Jun 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses standard shell utilities to perform local file discovery and keyword searches.
- Evidence: Employs
find,grep, andxargsinSKILL.mdto locate files containing strings related to onboarding and growth experiments (e.g., 'referral', 'experiment', 'sendgrid'). - [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it reads and processes local project files whose content could potentially contain malicious instructions.
- Ingestion points: Local project files including
.tsx,.jsx,.py, and.mdlocated via shell commands. - Boundary markers: Absent; there are no explicit delimiters or instructions to ignore content that mimics agent commands.
- Capability inventory:
Read,Bash,WebFetch,WebSearch,AskUserQuestion. - Sanitization: None; the skill directly processes the contents of searched files to generate a growth audit report.
- [SAFE]: The reconnaissance patterns are read-only, restricted to the local filesystem, and align perfectly with the skill's stated purpose of auditing growth infrastructure. All external references point to the author's official domains.
Audit Metadata