ClassifyCCP

Pass

Audited by Gen Agent Trust Hub on Jun 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes Python snippets and Bash commands to handle PDF parsing (via pymupdf) and spreadsheet manipulation (via openpyxl). These operations are local and triggered by user-provided files.- [PROMPT_INJECTION]: The skill ingests untrusted text from user-provided PDF documents, creating a surface for indirect prompt injection where embedded instructions could influence the classification logic.
  • Ingestion points: PDF content is extracted in Phase 1 and processed throughout the workflow.
  • Boundary markers: The instructions do not define clear delimiters or include prompts to ignore instructions within the source documents.
  • Capability inventory: The agent has capabilities for file system writes, local Python execution, and Bash command execution.
  • Sanitization: No validation or sanitization is performed on the text extracted from the PDF before it is analyzed.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 10, 2026, 02:41 PM
Security Audit — agent-trust-hub — ClassifyCCP