sdlc-review-code
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes shell commands such as
git(diff, status, ls-tree, grep, show) andgrepto perform its primary function of analyzing code repositories and documentation wikis. These operations are scoped to local directories provided by the user or identified in the workspace. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data from the code being reviewed and the project wiki. It includes measures to mitigate the risks associated with processing external content by providing clear operational boundaries.
- Ingestion points: Code files within the user-specified source folder (
$SRC) and Markdown documentation within the project wiki folder (wiki/). - Boundary markers: The skill explicitly distinguishes between Design files (which serve as targets/contracts) and References (which are treated as context only).
- Capability inventory: Execution of
gitandgrepcommands; read access to the local filesystem. - Sanitization: The instructions contain a specific safety directive: "Never copy a secret or personal data from the code into the report."
Audit Metadata