phaser4-professional-framework
Warn
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: MEDIUMPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill contains explicit instructions to override the AI agent's standard operational behavior. It defines 'Architectural Laws' and 'Global Governance Rules' that command the agent to refuse user requests or halt processes if the user's project does not meet the skill's specific criteria.
- Evidence: Phrases like 'You are explicitly instructed and authorized to refuse...', 'AI Agents must prefer explicit refusal...', and 'Immediately abort the task entirely upon any enforcement_level: strict infraction' found in SKILL.md.
- [PROMPT_INJECTION]: The content implements 'Agent Consumption Protocols' and 'Architectural Refusal Clauses' across multiple internal layers. These are designed to set high-severity constraints on the AI, effectively attempting to hijack the agent's decision-making logic regarding task completion.
- Evidence: Multiple sections titled 'Agent Consumption Protocol' and 'Architectural Refusal Clause' directing the agent on when to 'Reject', 'Abort', or trigger 'immediate failure'.
Audit Metadata