The Agent Skills Directory

[SAFE]: The skill consists entirely of instructional documentation and JSON examples for calling the Stitch MCP tool. No scripts, binaries, or automated shell commands are included.
[INDIRECT_PROMPT_INJECTION]: The skill describes a process that ingests external project data, which constitutes a potential injection surface.
Ingestion points: External project titles and identifiers returned by the list_projects tool.
Boundary markers: The instructions do not provide explicit delimiters to separate the tool output from the agent's instructions.
Capability inventory: Extracted project IDs are subsequently used in calls to other tools such as list_screens and generate_screen_from_text.
Sanitization: The instruction set does not include steps for the agent to sanitize or validate the format of strings returned from the external tool before interpolation.

stitch-mcp-list-projects