skills/mateonunez/skills/caveman/Gen Agent Trust Hub

caveman

Pass

Audited by Gen Agent Trust Hub on May 12, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill uses persistence instructions such as "ACTIVE EVERY RESPONSE once triggered" and "No revert after many turns" to ensure the 'caveman' persona overrides standard model behavior and conversational drift. While intended for stylistic consistency and token efficiency, these patterns resemble override techniques used in prompt injections to maintain control over the AI's output state.
  • [SAFE]: The skill includes an 'Auto-Clarity Exception' that explicitly instructs the agent to drop the persona for security warnings and confirmations of irreversible actions, which serves as a safety guardrail.
Audit Metadata
Risk Level
SAFE
Analyzed
May 12, 2026, 04:03 PM
Security Audit — agent-trust-hub — caveman