codebase-audit
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data from a target codebase using tools like Read, Grep, and Glob. This introduces a surface for indirect prompt injection where malicious instructions embedded within the audited code could attempt to subvert the agent's audit logic.
- Ingestion points: SKILL.md instructs the agent to read and search through codebase files.
- Boundary markers: Absent; there are no instructions to ignore or delimit prompt-like content within files.
- Capability inventory: The agent has access to Bash for command execution and WebFetch for network research.
- Sanitization: Absent; the skill performs raw text analysis of the files.
- [COMMAND_EXECUTION]: The skill orchestrates the execution of system-level audit utilities (e.g., trufflehog, npm audit, pip-audit) using the Bash tool. While these are standard security practices, the use of shell tools in an environment where the agent's reasoning is influenced by untrusted codebase content carries an inherent risk.
Audit Metadata