new-case

Pass

Audited by Gen Agent Trust Hub on Apr 24, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to use the shell command date "+%Y-%m-%d" to ensure document creation dates and timelines are accurate. This is a benign use of system tools for metadata purposes.
  • [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it is designed to ingest and process untrusted external data.
  • Ingestion points: The agent scans directory contents and reads various file types (chat records, contracts, legal documents) to extract metadata as specified in references/extraction-rules.md.
  • Boundary markers: The instructions do not define clear delimiters or system-level warnings to ignore embedded instructions within the processed legal materials.
  • Capability inventory: The skill can create directories, move files, and generate new Markdown and YAML files based on the analyzed content.
  • Sanitization: There are no specific sanitization or validation steps described to filter potentially malicious text embedded in the ingested materials.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 24, 2026, 09:41 AM