rails-conventions
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local shell commands including
bin/rubocop,bundle exec rubocop, andgit diffto perform code quality checks. These are standard development tools and the command arguments are statically defined. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data by reading repository code through
git diffand possesses shell execution capabilities (RuboCop, test suites). This creates a vulnerability surface where malicious content embedded in the codebase could theoretically attempt to influence the agent's behavior. - Ingestion points: Reads the output of
git diff --cached -- '*.rb'(SKILL.md). - Boundary markers: Absent; the agent scans the raw diff output for specific patterns without explicit isolation markers.
- Capability inventory: Execution of
bin/rubocop,bundle exec rubocop,git diff, and project test suites (SKILL.md). - Sanitization: None; the raw output of repository commands is processed directly by the agent.
Audit Metadata