analyze
Pass
Audited by Gen Agent Trust Hub on Jul 29, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a secure, read-only approach to repository analysis. It includes strong internal guardrails, such as a strict non-negotiable contract against editing files, a requirement for explicit evidence-vs-inference distinctions, and instructions to avoid speculative or judgmental language.
- [PROMPT_INJECTION]: The skill demonstrates an inherent surface for indirect prompt injection as its core functionality involves reading and interpreting external, untrusted codebase data.
- Ingestion points: Repository files accessed via Read, Grep, Glob, and read-only Bash tools.
- Boundary markers: The instructions do not define specific delimiters to isolate file content from the agent's instruction stream.
- Capability inventory: Spawning sub-agents (Agent tool) for parallel exploration and executing read-only shell commands.
- Sanitization: No explicit sanitization or escaping of ingested file content is mandated in the skill logic.
Audit Metadata