kaizen-codebase
Pass
Audited by Gen Agent Trust Hub on Sep 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands for repository management and change verification.
- Uses
gitfor worktree isolation, fetching remote updates, and managing logical commits. - Executes local test suites, linters, and type-checking tools to ensure code behavior is preserved before opening a pull request.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the local repository, creating a surface where malicious instructions in codebase files could influence agent behavior.
- Ingestion points: Reads
CONTEXT.md, Architecture Decision Records (ADRs), and source code files to identify improvement candidates (SKILL.md). - Boundary markers: The instructions do not specify the use of delimiters or explicit boundary markers when interpolating codebase content into the subagent context.
- Capability inventory: The skill has the ability to modify the filesystem, execute shell commands (via tests), and interact with GitHub via CLI tools (SKILL.md).
- Sanitization: No mention is made of filtering, escaping, or validating the content read from the repository files before processing it.
Audit Metadata