docs-finalize-and-commit
Pass
Audited by Gen Agent Trust Hub on Mar 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill detects and executes build scripts defined in the repository (e.g., scripts in
package.json,Makefile, or framework-specific CLIs) during Gate 5 (Build Verification). This is a standard and necessary behavior to ensure documentation integrity before committing. - [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it reads and analyzes untrusted repository content (documentation files and configuration files) to infer project conventions.
- Ingestion points: Documentation files (
.md,.mdx) and framework config files (e.g.,docusaurus.config.js,mkdocs.yml). - Boundary markers: None are explicitly used when reading files; the agent is instructed to sample headers and deep-read content.
- Capability inventory: Filesystem writes (
Auto-Fix), shell command execution (Build Verification), andgitoperations (staging and committing). - Sanitization: There is no explicit sanitization of documentation content before the agent processes it for convention discovery.
- [SAFE]: The skill incorporates strong safety measures in Gate 0 (Working Set Validation), which isolates changes to the current session and strictly forbids the use of commands that would discard or revert uncommitted changes in other files (e.g.,
git restore,git reset --hard).
Audit Metadata