plan-writing

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is a set of markdown-based instructions for task planning. It defines principles for breaking down work, verifying outcomes, and naming plan files within a project. It does not contain executable code, hidden payloads, or suspicious network requests.
  • [COMMAND_EXECUTION]: The skill provides examples of commands (e.g., npm install, npx create-next-app) and script names (e.g., security_scan.py, api_validator.py) that an agent should include in its generated plans. These are contextually appropriate for developer-oriented task planning and do not represent a direct security threat within the skill itself.
  • [PROMPT_INJECTION]: Indirect prompt injection surface detected as the skill processes user-defined tasks to generate plans.
  • Ingestion points: User-provided task descriptions are used to derive {task-slug} and plan content.
  • Boundary markers: None defined to isolate user input from the generated plan structure.
  • Capability inventory: The skill suggests generating plans that execute shell commands and local python scripts.
  • Sanitization: No explicit instructions are provided for sanitizing user-provided task names or descriptions before they are used in file naming or plan generation.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 01:14 AM