docs-finalize-and-commit
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes various shell commands to perform its tasks. This includes standard Git operations (
git status,git diff,git log,git add,git commit,git push) and project-specific build commands. - Build Command Execution: In Gate 5, the skill is instructed to detect and execute build commands from the project configuration (e.g.,
package.jsonscripts,Makefile, or framework-specific CLIs). This involves executing arbitrary commands defined within the local workspace. - [INDIRECT_PROMPT_INJECTION]: The skill's core functionality involves reading and analyzing existing documentation files and source code to "discover conventions" (tone, terminology, and structure).
- Ingestion points: Gate 1 reads the frontmatter and the first 30 lines of 10–15 documentation files. Gate 2 reads source code diffs from the working tree.
- Boundary markers: The instructions do not specify any boundary markers or delimiters to prevent the agent from being influenced by instructions hidden within the sampled documentation or source code.
- Capability inventory: The skill possesses significant capabilities, including file modification (Auto-Fix), staging/committing changes via
git, and executing external build scripts. - Sanitization: There is no explicit sanitization or filtering of the content sampled from the repository files before it is used to build the "Convention Reference" that guides subsequent agent actions.
Audit Metadata