Skills
Navigate
skills/shinpr/claude-code-workflows/recipe-front-review/Gen Agent Trust Hub

recipe-front-review

Pass

Audited by Gen Agent Trust Hub on Nov 8, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from the repository's code and design documents as input for sub-agent analysis.
  • Ingestion points: implementationFiles (source code and configuration) and Design Doc ($ARGUMENTS) in Step 1.
  • Boundary markers: Prompt interpolation uses structured list formatting for file paths and contents.
  • Capability inventory: The skill can modify local files and design documentation via the task-executor-frontend and technical-designer-frontend sub-agents.
  • Sanitization: The risk is mitigated by a mandatory human-in-the-loop approval step (Step 4) and specialized review resolution gates that validate sub-agent outputs.
  • [COMMAND_EXECUTION]: The skill uses local repository state (git merge base, working tree, untracked files) to identify the set of files to be reviewed. This is a standard function for development-oriented automation tools.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 8, 2026, 07:05 PM