tech-due-diligence
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external codebases which may contain adversarial content designed to influence the agent's behavior or report generation.
- Ingestion points: SKILL.md (Workflow Step 1) accepts GitHub URLs and local directory paths for evaluation.
- Boundary markers: No specific delimiters or 'ignore' instructions are provided to separate the codebase content from the agent's primary instructions.
- Capability inventory: The skill has access to tools including Bash, Read, Glob, and Grep.
- Sanitization: No sanitization or validation logic is specified for the data ingested from target codebases.
- [COMMAND_EXECUTION]: The skill uses the Bash tool to execute shell commands for repository metrics and history analysis.
- Evidence: The investigation protocol (Phase 1 and Phase 8) utilizes standard utilities like 'find', 'wc', and 'git shortlog'.
- [EXTERNAL_DOWNLOADS]: The skill clones codebases from external URLs to perform its primary function.
- Evidence: SKILL.md instructs the agent to clone GitHub URLs as the first step of the workflow.
Audit Metadata