research-staff-meeting

Pass

Audited by Gen Agent Trust Hub on Jun 23, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: The skill defines a benign workflow for documenting participation in research meetings using markdown files.
  • [NO_CODE]: The skill consists entirely of markdown instructions and metadata. It does not include any executable scripts, shell commands, or package dependencies.
  • [SAFE]: Indirect Prompt Injection Analysis:
  • Ingestion points: The agent is instructed to process meeting content during the PREPARE and CHALLENGE phases.
  • Boundary markers: None specified in the instructions.
  • Capability inventory: File-write operations restricted to judgments/ and critiques/ directories.
  • Sanitization: No explicit input sanitization is mentioned, though the output is limited to markdown documentation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 23, 2026, 07:11 AM
Security Audit — agent-trust-hub — research-staff-meeting