project-wiki
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes
git submodule update,git commit, andgit pushshell commands to manage repository content and synchronize generated documentation. - [EXTERNAL_DOWNLOADS]: The skill triggers downloads of external code from remote sources specified in the project's
.gitmodulesfile during the execution of submodule update commands. - [INDIRECT_PROMPT_INJECTION]:
- Ingestion points: The skill performs a comprehensive analysis of the repository workspace (SKILL.md), including source files, entry points, configuration files (existing-code-docs.md), and git commit history.
- Boundary markers: The instructions do not define explicit delimiters or instructions to ignore embedded prompts within the analyzed files.
- Capability inventory: The agent can read all workspace files, write markdown documentation to the local system, and execute git commands (sync-code-docs.md, existing-code-docs.md).
- Sanitization: The instructions focus on preventing credential leakage through placeholders but do not address the filtering of malicious natural language instructions found in project data.
Audit Metadata