review-merge-readiness
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- PROMPT_INJECTION (LOW): Indirect Prompt Injection Surface. The skill ingests untrusted data from user-provided implementation descriptions and plan documents which could contain instructions intended to bias the review.
- Ingestion points: The
code-reviewer.mdtemplate interpolates{DESCRIPTION}and{PLAN_REFERENCE}directly into the prompt instructions for the sub-agent. - Boundary markers: Absent. The untrusted content is not delimited by markers like
---or XML tags, nor does it include warnings to the AI to ignore instructions within the data. - Capability inventory: The agent has the capability to execute
git diffand provide a production readiness verdict. - Sanitization: Absent. No escaping or validation is performed on the input variables.
- COMMAND_EXECUTION (SAFE): The skill utilizes standard
gitcommands (git merge-base,git rev-parse,git diff) to determine the scope of changes and extract code evidence. These operations are essential for the primary purpose of code review and are restricted to read-only git metadata/diff operations.
Audit Metadata