The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to run the command scripts/stitch_setup.py check to verify the environment. This represents direct execution of code on the local system.
[DYNAMIC_EXECUTION]: The setup process involves running a Python script (stitch_setup.py). Since the source code for this script is not included in the skill package files for inspection, its runtime behavior cannot be verified.
[INDIRECT_PROMPT_INJECTION]: The skill determines which workflow to execute based on untrusted user requests. While the instructions include several safety checks, this logic is inherently susceptible to manipulation via embedded instructions in user data.
Ingestion points: Natural language user intents and design task descriptions processed in SKILL.md.
Boundary markers: The skill utilizes 'Rules' and 'Workflow' steps that emphasize explicit user authorization and evidence-based confirmation.
Capability inventory: Includes executing local scripts, routing to design system managers, and performing remote project operations.
Sanitization: The skill mandates that evidence must be verified and prohibits using guessed values or implicitly expanding authorization.

stitch-design-use