claude-gif-create
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes various system commands to manage the animation project, including Node.js tools (
node,npx,npm) and local utility scripts. It runsgif_frames.py,gif_convert.sh, andgif_optimize.pylocated within the skill's own internal directory structure (~/.claude/skills/claude-gif/scripts/) to process image sequences into optimized GIF files. - [DYNAMIC_EXECUTION]: The skill generates React source code for an animation (
GifComposition.tsx) at runtime based on the user's request. This generated code is then executed by the Remotion rendering engine to produce PNG frames. While this is the intended functionality of the skill, it involves the dynamic creation and execution of code. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied text and animation requirements which are interpolated into the React component code by the agent.
- Ingestion points: User-provided descriptions of the animation type, content, and colors in SKILL.md.
- Boundary markers: The instructions do not define specific delimiters or guards for user-supplied strings within the generated JSX.
- Capability inventory: The agent has access to
Bash,Read,Write, andEdittools, allowing it to modify the local environment and execute scripts. - Sanitization: There are no explicit instructions for sanitizing user-provided strings before they are written into the React component file.
Audit Metadata