code-review-secrets-detection
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and process data from untrusted external sources, such as third-party Git repositories, exposed
.gitdirectories, and CI/CD configuration files. - Ingestion points: Untrusted data is processed when the agent uses tools like
trufflehog,gitleaks, andgit-dumperon external repositories or web servers (SKILL.md). - Capability inventory: The agent is tasked with executing command-line utilities and performing network validation calls (e.g.,
aws sts get-caller-identity) using credentials discovered during the scan. - Boundary markers: The skill does not provide specific markers or instructions to isolate the scanned content from the agent's control flow, creating a risk that malicious instructions inside the scanned code could influence agent behavior.
- Sanitization: No sanitization or validation mechanisms are described to filter potentially malicious payloads within the untrusted files before they are read by the agent.
Audit Metadata