implement

Pass

Audited by Gen Agent Trust Hub on Mar 30, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes instructions and context from external project files such as agents.md and implementation plans in the .ai directory, creating an indirect prompt injection surface. * Ingestion points: Reads agents.md, .ai/implementation-plan.md, .ai/issue-analysis.md, and .ai/plan-review-report.md (SKILL.md). * Boundary markers: None identified in the skill instructions. * Capability inventory: The skill writes code and tests to the file system and executes shell commands for testing and deployment (SKILL.md). * Sanitization: No validation or sanitization of content from these project files is performed.
  • [COMMAND_EXECUTION]: The skill is designed to execute shell commands for testing and deployment as defined in the project's agents.md file.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 30, 2026, 12:21 AM