The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local bash script, scripts/fetch-stitch.sh, to download design assets. It also instructs the agent to run npm install to set up the development environment.
[EXTERNAL_DOWNLOADS]: The skill uses curl within its internal bash script to fetch HTML content from external Stitch design URLs (Google Cloud Storage). These downloads are necessary for the skill's primary function of design conversion.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted HTML and metadata from external designs. It lacks explicit instructions for the agent to sanitize this input or use protective boundary markers, which presents a surface for indirect prompt injection where malicious text in a design file could influence agent behavior.
Ingestion points: SKILL.md and scripts/fetch-stitch.sh define a workflow for retrieving and parsing external HTML source code.
Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are present in the prompt templates.
Capability inventory: The skill possesses the ability to execute shell commands (bash, npm) and write files to the local file system.
Sanitization: No explicit sanitization or filtering logic is defined for the external design content before processing.

stitch-ui-vue-bootstrap-components