conventions
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions for setting up a git hook that executes a local Node.js script (
cli.mjs) and potentially external CLI tools likeclaudeorcodexfor linting.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes commit messages, which are external untrusted inputs. These inputs are used as data for the linter and may be passed to LLMs, presenting an indirect prompt injection surface.\n - Ingestion points: Git commit message file via the
$1argument in the hook script specified inSKILL.md.\n - Boundary markers: None identified in the provided configuration.\n
- Capability inventory: Execution of
node cli.mjs,claude -p, andcodex execas defined inconventions.yml.\n - Sanitization: No sanitization or filtering of the commit message content is described before it is passed to external tools.
Audit Metadata