template-meta-prompt-creation

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION] (LOW): The skill defines an attack surface for indirect prompt injection through its meta-prompt templates. 1. Ingestion points: The example meta-prompt processes untrusted user data via the $ARGUMENTS variable (SKILL.md). 2. Boundary markers: The templates do not include delimiters or instructions to ignore embedded commands within the processed data. 3. Capability inventory: The suggested prompts utilize high-privilege tools including Write, Edit, WebFetch, and Task (SKILL.md). 4. Sanitization: No sanitization or validation of external content is suggested or implemented.
  • [EXTERNAL_DOWNLOADS] (SAFE): Documentation links point exclusively to trusted sources (docs.anthropic.com).
  • [COMMAND_EXECUTION] (SAFE): The skill's own allowed tools (Read, Grep, Glob) are restricted and appropriate for its stated purpose.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:28 PM