spec-miner
Pass
Audited by Gen Agent Trust Hub on Jun 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's logic is entirely focused on analyzing and documenting local source code using standard utilities (Glob, Grep, Read). No suspicious commands or hidden behaviors were detected across the skill files.
- [COMMAND_EXECUTION]: The skill utilizes Bash, Grep, and Glob for local code analysis. This is appropriate for its stated purpose of reverse-engineering undocumented or legacy systems, and the examples provided show legitimate usage for finding entry points, routes, and configuration patterns.
- [DATA_EXFILTRATION]: There are no network operations or attempts to exfiltrate data. While the skill identifies configuration and environment variables, it does so to document system architecture rather than to harvest or transmit secrets.
- [PROMPT_INJECTION]: The skill processes untrusted data in the form of existing source code, which represents a potential surface for Indirect Prompt Injection. However, the risk is minimized by strict instructional constraints that require the agent to ground all observations in code evidence and use specific output formats (EARS syntax).
- Ingestion points: Local files accessed via Glob, Grep, and Read tools during exploration.
- Boundary markers: The output is strictly formatted according to markdown templates and the EARS requirement syntax.
- Capability inventory: File system access (Read, Grep, Glob) and Bash for local command execution.
- Sanitization: The instructions explicitly require the agent to distinguish between facts and inferences and to document uncertainties.
Audit Metadata