jianying-edit

Pass

Audited by Gen Agent Trust Hub on Sep 19, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill relies on a local binary named jianying to perform media probing, project creation, and draft editing. This execution is controlled through explicit doctor and capabilities checks to ensure the environment is supported before any operations are attempted.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external media files and existing project drafts, which represents a potential injection surface from file metadata.
  • Ingestion points: Media files are analyzed via jianying media probe, and existing drafts are loaded for editing.
  • Boundary markers: The workflow utilizes strict JSON schemas (jianying-job/v2) and requires structural verification (project verify) to ensure project integrity.
  • Capability inventory: CLI tools are used for specific, scoped tasks such as reading media metadata and writing to isolated draft copies.
  • Sanitization: The skill explicitly forbids using inferred values, requiring all media facts (durations, dimensions, hashes) to be sourced directly from tool outputs and validated against a plan before execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 19, 2026, 11:55 PM
Security Audit — agent-trust-hub — jianying-edit