requesting-code-review
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill uses standard, benign git commands (
git rev-parse,git log,git diff,git show) to inspect code changes for review purposes. These are necessary for the skill's primary function and do not target sensitive files or system configurations. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided descriptions and plans (
{DESCRIPTION},{PLAN_OR_REQUIREMENTS}) by interpolating them into a subagent prompt. While this creates a surface for indirect injection, the risk is minimal as the subagent is explicitly instructed to be 'read-only' and is used for internal analysis rather than high-privilege actions. The template includes clear boundary markers and specific instructions on how to handle the data. - [DYNAMIC_CONTEXT_INJECTION]: The skill provides examples of shell commands used to populate variables (
BASE_SHA,HEAD_SHA). These commands are benign and standard for development workflows. There are no instances of the specific!commandsyntax that triggers silent execution at skill load time.
Audit Metadata