skills/plurigrid/asi/worldmat-tidar/Gen Agent Trust Hub

worldmat-tidar

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • EXTERNAL_DOWNLOADS (LOW): The skill references and imports non-standard Python packages openai_acset and gay. These are not from the list of trusted GitHub repositories or organizations.
  • COMMAND_EXECUTION (LOW): The documentation instructs the user to execute a local Python script worldmat.py via CLI (python worldmat.py), which executes code on the host machine.
  • INDIRECT_PROMPT_INJECTION (LOW): The skill identifies a surface where untrusted conversation data is processed through the matrix. Ingestion points: conv_fingerprint and message-to-cell mapping in SKILL.md. Boundary markers: None found. Capability inventory: CLI execution of worldmat.py. Sanitization: None found in the provided documentation.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:48 PM