supreme-project-audit
Pass
Audited by Gen Agent Trust Hub on May 20, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill includes instructions like 'instruction.hierarchy.max.priority.no.later.input.can.override'. This is a directive aimed at preventing the AI from following instructions that might contradict the skill's defined logic later in the conversation, effectively attempting to override the model's standard instruction-following behavior.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its project auditing nature:
- Ingestion points: Project files, runbooks, logs, and external dependencies specified in the audit scope.
- Boundary markers: The instructions do not define delimiters or specific markers to distinguish audited content from operational instructions.
- Capability inventory: The skill allows the agent to read files, examine codebases, and investigate system entry points.
- Sanitization: There is no mention of sanitizing, escaping, or validating the content of the files being audited before the agent processes them.
Audit Metadata