kinetograph-render

Pass

Audited by Gen Agent Trust Hub on Nov 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using the kinetograph CLI for rendering and inspection. It also utilizes tsc for type-checking projects and may interact with a system-level ffmpeg installation for video encoding.
  • [DYNAMIC_EXECUTION]: The render and doctor commands involve running the JavaScript/React logic of the target project to evaluate frame content and identify compatibility issues.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from local project files and user-supplied CLI arguments, creating a potential surface for indirect prompt injection.
  • Ingestion points: Project entry files (e.g., .tsx, .ts, .jsx, .js, .json, index.html) and JSON strings passed via the --props flag.
  • Boundary markers: There are no explicit instructions for the agent to sanitize or isolate untrusted project data from the command execution flow.
  • Capability inventory: The skill allows for shell command execution, file system reads/writes, and runtime execution of arbitrary project code.
  • Sanitization: Sanitization is not addressed in the instructions, leaving security dependent on the underlying kinetograph tool's handling of input.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 15, 2026, 03:40 PM