The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute internal logic and interact with specialized CLI tools.
It runs scripts/analyze-images.ts using npx tsx to perform product and model image analysis.
It invokes renoise-cli.mjs for uploading materials and submitting video generation tasks.
[EXTERNAL_DOWNLOADS]: The skill performs routine external data retrieval and dependency management.
It executes npm install within the skill directory to ensure required Node.js libraries are available for the analysis script.
It employs the WebFetch tool to retrieve product details from user-provided URLs, enhancing the context for script generation.
[PROMPT_INJECTION]: The skill's ingestion of external web content and images presents a surface for indirect prompt injection (Category 8).
Ingestion points: Product images, model reference photos, and external product detail pages fetched via WebFetch.
Boundary markers: The image analysis script enforces a structured JSON response through its hardcoded system prompt, which limits the influence of adversarial content.
Capability inventory: The skill has access to local file reading and Bash command execution to support its core functions.
Sanitization: External content is processed by the analysis model without pre-validation, relying on the model's instructions and structured output requirements for safety.

tiktok-content-maker