deliver-review
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes and analyzes diffs and code from arbitrary repository branches, which could contain instructions meant to influence the agent.
- Ingestion points: The reviewer reads diffs and repository content from target branches as part of the review/repair loop (SKILL.md).
- Boundary markers: There are no specified delimiters or instructions to ignore embedded prompts within the code being analyzed.
- Capability inventory: The skill can execute "baseline" tests, create/modify git branches and worktrees, and perform automated commits via an "implementer."
- Sanitization: No sanitization or validation of the repository content is mentioned before it is processed by the agent.
- [DYNAMIC_EXECUTION]: The skill is instructed to "Run the baseline" and "follow the shared review loop," which typically entails executing test suites or scripts defined within the repository under review.
- [COMMAND_EXECUTION]: The skill interacts with an "adapter" to perform git operations, manage worktrees, and maintain temporary workspaces, implying the execution of shell commands.
Audit Metadata