melech-diff-minimizer
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external data from Pull Request metadata, descriptions, and code diffs, which could potentially contain adversarial instructions.
- Ingestion points: Resolves comparison targets from PR metadata and repository conventions; reconciles PR descriptions and tests in Workflow Step 1; inspects whole diffs in Workflow Step 2.
- Boundary markers: The instructions do not define specific delimiters or instructions to ignore potential commands embedded within the PR content.
- Capability inventory: The skill instructs the agent to execute repository verification checks and tests in Workflow Step 5.
- Sanitization: No explicit sanitization or filtering of PR metadata or descriptions is specified before the agent processes the content.
- [COMMAND_EXECUTION]: The workflow requires the agent to run verification checks and tests within the repository environment to confirm that behavioral equivalence is maintained after the diff is minimized.
Audit Metadata