Requesting Code Review
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes
git rev-parseinSKILL.mdandgit diffincode-reviewer.mdto identify commit hashes and display code changes. These are standard operations for development workflows. - [INDIRECT_PROMPT_INJECTION]: The skill facilitates the review of untrusted code which could contain malicious instructions designed to manipulate the subagent.
- Ingestion points: The
code-reviewer.mdtemplate ingests external content from git diff outputs and placeholders like{PLAN_OR_REQUIREMENTS}. - Boundary markers: Absent; the instructions in
code-reviewer.mddo not include explicit delimiters or warnings to disregard instructions found within the code changes. - Capability inventory: Subprocess calls to git for repository inspection are present in both
SKILL.mdandcode-reviewer.md. - Sanitization: Absent; the content processed from git and user placeholders is interpolated into the agent context without escaping or validation.
Audit Metadata