implement-with-design

Pass

Audited by Gen Agent Trust Hub on May 14, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted content from local plan files and documentation to drive its implementation loop.
  • Ingestion points: Plan files (plan.md), UI rules (UI-RULES.md), design tokens (DESIGN.md), and library caches (.library/) as specified in SKILL.md.
  • Boundary markers: The skill does not define specific markers to isolate instructions from data within these files.
  • Capability inventory: The agent can write source code, create git commits, execute package manager commands for versioning, and perform web searches via Exa.
  • Sanitization: No input sanitization or verification mechanism is described for the content of the plan files.
  • [COMMAND_EXECUTION]: The skill instructs the agent to use various system commands for project management and environment discovery.
  • Uses git status, git add, and git commit to manage the working tree.
  • Uses package manager CLI tools (e.g., npm view, pip index, cargo search) to verify dependency versions.
  • [EXTERNAL_DOWNLOADS]: The skill fetches external data to support the implementation process.
  • Uses Exa tools (get_code_context_exa, web_search_exa) and Context7 for library research and documentation.
  • Queries official package registries for version information.
Audit Metadata
Risk Level
SAFE
Analyzed
May 14, 2026, 02:54 PM
Security Audit — agent-trust-hub — implement-with-design