The Agent Skills Directory

[SAFE]: The skill implements a persistent knowledge mechanism by reading from and appending to references/learnings.md to store platform-specific insights and workarounds discovered during use.
[EXTERNAL_DOWNLOADS]: Provides references to official vendor resources, including the tl;dv API at pasta.tldv.io and the official MCP server repository on GitHub for advanced tool integration.
[COMMAND_EXECUTION]: Documentation includes a suggested command (npx skills add) for users to install related skills from the same author repository.
[PROMPT_INJECTION]: The skill processes user-supplied meeting details and platform questions, creating a surface for indirect prompt injection through ingested meeting context and data.
Ingestion points: SKILL.md (user queries) and references/learnings.md.
Boundary markers: Absent.
Capability inventory: Agent instructions to read from and write to references/learnings.md based on interaction outcomes.
Sanitization: Absent.

sales-tldv