source-code-review
Pass
Audited by Gen Agent Trust Hub on Jul 6, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted application source code which may contain malicious instructions.
- Ingestion points: Source code files and configuration files (e.g., .env, settings.py, Web.config) read from the target source path.
- Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded commands within the analyzed data, although it utilizes subagents for the initial scanning phase.
- Capability inventory: The agent can read local files, use glob patterns, and write findings to markdown files in the engagement directory.
- Sanitization: There is no evidence of sanitization or escaping of the content extracted from the source files before it is analyzed by the agent.
- [DATA_EXFILTRATION]: The skill's core purpose is the identification and collection of sensitive data including hardcoded passwords, API keys, database connection strings, and private keys. While the instructions direct the agent to store these findings locally within an engagement directory, the automated aggregation of such secrets represents a data exposure risk if the environment is compromised.
Audit Metadata