consciousness-council
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSNO_CODE
Full Analysis
- [SAFE]: The skill consists entirely of natural language instructions and structured templates. It contains no executable scripts, binaries, or platform-specific tool configurations, ensuring no risk of unauthorized system operations or command execution.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a framework for ingesting and processing user-supplied objectives. It includes explicit defensive instructions to distinguish factual claims from assumptions and to verify evidence against external sources, which helps mitigate the risk of the agent adopting malicious instructions embedded in processed data.
- [EXTERNAL_DOWNLOADS]: The skill contains links to academic preprints on ArXiv and the author's official websites for attribution. These are informational references and do not involve automated retrieval or execution of remote content.
- [NO_CODE]: There are no Python or Node.js packages, shell scripts, or external dependencies associated with this skill, eliminating vectors for remote code execution, persistence mechanisms, or privilege escalation.
Audit Metadata