execute-plan
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes a vendor-provided Python script,
scripts/quality_gates.py, which executes localgitcommands viasubprocess.runto perform repository analysis, export reachability checks, and diff hygiene. These operations are performed using a secure execution method (list-based arguments without a shell) and are essential for the skill's primary function of verifying code integrity. - [INDIRECT_PROMPT_INJECTION]: The skill processes external implementation plans (
docs/plans/*.md) which serve as an ingestion point for potentially untrusted data. To mitigate injection risks, the instructions mandate strict file-ownership boundaries for sub-agents, manual review of generated diffs, and the use of automated quality gates to ensure code follows established invariants and safety rules. - [SAFE]: The
references/build-rules.mdfile defines a comprehensive set of security best practices that agents must follow, including requirements for input validation, constant-time comparisons for secrets, secure transaction handling, and the prohibition of stubs or placeholders in production code. These rules act as a defensive framework for all code generated by the skill.
Audit Metadata