repo-scan
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill operates by ingesting and analyzing untrusted source code and metadata from external repositories, which presents a surface for indirect prompt injection attacks.
- Ingestion points: File contents, commit messages, and repository metadata are ingested via the
ReadandBashtools during the classification and hotspot analysis phases (e.g., inguides/codebase-security-audit-workflow.md). - Boundary markers: The skill does not employ explicit delimiters or system instructions to ignore embedded prompts within the ingested repository data.
- Capability inventory: The skill possesses high-privilege capabilities including
Bashfor command execution,WebFetchfor network access, and theAgenttool for task automation. - Sanitization: No sanitization or filtering of the repository content is performed before the agent processes the findings, potentially allowing embedded malicious instructions to affect the agent's logic.
- [COMMAND_EXECUTION]: The skill methodology utilizes the
Bashtool to execute a wide array of security-focused CLI utilities (semgrep, trivy, trufflehog, gitleaks). These executions are documented and intended for the skill's primary purpose of security auditing. - [EXTERNAL_DOWNLOADS]: The documentation references and provides setup guides for various security tools available through official package registries (pip, npm, cargo, etc.) and GitHub Actions. All external downloads and service references target trusted organizations or well-known services, such as Aqua Security and GitHub.
Audit Metadata