raven-investigate
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCREDENTIALS_UNSAFECOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted data from the codebase being audited, including source code and configuration files. It lacks explicit boundary markers or sanitization for this content, which could allow malicious comments or data within an audited project to influence the agent's behavior. This is a common vulnerability surface for tools that analyze external content.
- [CREDENTIALS_UNSAFE]: The tool is specifically designed to scan for and read sensitive file paths, including
.envfiles,.aws/credentials, and SSH keys (e.g.,~/.ssh/id_rsa). This behavior is the primary intended function of the 'Secrets & Credentials' audit beat. The risk is mitigated by explicit instructions to describe findings (e.g., file path and line number) without reproducing the actual secret values in the generated report. - [COMMAND_EXECUTION]: The skill utilizes system commands such as
ls,find, and standard package manager audit tools (e.g.,npm audit,pip audit,cargo audit) to perform its analysis. These are legitimate tools for a security auditor and are executed in a controlled manner based on the detected tech stack.
Audit Metadata