thermo-fix
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests potentially untrusted file content from the diff scope to perform automated fixes, creating a risk surface for indirect instructions. * Ingestion point: Files in diffScope. * Boundary markers: None specified to separate code content from instructions. * Capability inventory: File writing, shell command execution, and git operations. * Sanitization: None mentioned for ingested code content.
- [COMMAND_EXECUTION]: The skill executes a user-supplied buildCommand to verify fixes, which involves shell command invocation in the local development environment.
Audit Metadata