security-threat-model
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an inherent attack surface common to security analysis tools that ingest and process untrusted repository source code. Maliciously crafted files within a scanned codebase could theoretically attempt to influence the agent's output.
- Ingestion points: The tool is designed to read and summarize entire codebases, including entry points, configurations, and logic flows, as specified in SKILL.md and references/prompt-template.md.
- Boundary markers: The instructions provide clear directives to ground all claims in repository evidence and to redact secrets, which serves as a functional control, though it lacks formal technical delimiters for data isolation.
- Capability inventory: The agent is authorized to read local files (potentially using tools like ripgrep) and write a summary threat model report to a Markdown file on the local file system.
- Sanitization: The instructions do not define specific content sanitization or escaping mechanisms for the source code data before it is processed by the model.
Audit Metadata