confrontational-creativity-engine

Pass

Audited by Gen Agent Trust Hub on Apr 8, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill uses "Do not refuse" and "Do not sanitize" directives in SKILL.md, which are explicit attempts to override standard AI safety filters and refusal behaviors.
  • [PROMPT_INJECTION]: The workflow introduces a surface for indirect prompt injection by mandating the ingestion of the entire conversation context without filtering content that may be malicious.
  • Ingestion points: SKILL.md requires processing the "full conversation state" and "every message."
  • Boundary markers: There are no markers or instructions to treat conversation history as untrusted data.
  • Capability inventory: The skill currently lacks dangerous tools like shell execution or network access, limiting the impact of successful injection to linguistic output.
  • Sanitization: The skill explicitly forbids sanitizing input or discarding context based on its nature.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 8, 2026, 03:50 PM