crg-req-to-design

Pass

Audited by Gen Agent Trust Hub on May 24, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection Surface. The skill processes untrusted hardware requirement data from Excel and CSV files, which could contain malicious instructions designed to influence the agent or downstream tools.
  • Ingestion points: Data enters the system via the ReqTableParser in scripts/req_parser.py which reads external files provided as input.
  • Boundary markers: The skill does not implement boundary markers or instructions to the agent to ignore potentially malicious embedded content within the table cells (e.g., in 'notes' or 'signal' columns).
  • Capability inventory: The skill performs file read and write operations, generating design tables and reports that are intended for further automated processing.
  • Sanitization: There is no evidence of sanitization or content validation for the data extracted from the input files.
  • Multi-step chains: The skill's primary purpose is to generate intermediate files for another skill (cr_tree_diag_gen), creating a chain where malicious data injected at the start could be propagated and executed by subsequent tools in the pipeline.
Audit Metadata
Risk Level
SAFE
Analyzed
May 24, 2026, 07:02 AM
Security Audit — agent-trust-hub — crg-req-to-design