fiftyone-develop-plugin

Pass

Audited by Gen Agent Trust Hub on Oct 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to transform user prompts into plugin code and execute development commands, which creates a surface for indirect prompt injection.
  • Ingestion points: User requirements for plugin functionality and UI design provided as natural language prompts in SKILL.md.
  • Boundary markers: Directives in SKILL.md instruct the agent to ask clarifying questions and present a plan for user approval before generating any code.
  • Capability inventory: The skill has the capability to write files, clone repositories via git clone, and execute commands such as npm, vite, and pytest as described in SKILL.md, HYBRID-PLUGINS.md, and PYTHON-OPERATOR.md.
  • Sanitization: No explicit sanitization or filtering of user-provided content is mentioned before it is incorporated into the generated code.
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions to download reference implementations from a remote repository.
  • Evidence: SKILL.md includes commands to clone the official plugin repository from github.com/voxel51/fiftyone-plugins.git for pattern reference.
  • [COMMAND_EXECUTION]: The skill utilizes various system and development tools to manage the plugin lifecycle.
  • Evidence: Instructions involve the use of git, npm, npx playwright, pytest, and ln -s for environment setup, building, and testing, as seen in SKILL.md and HYBRID-PLUGINS.md.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 17, 2026, 07:41 AM