The Agent Skills Directory

[PROMPT_INJECTION]: The skill handles untrusted data from project work items and comments, creating a potential surface for indirect prompt injection.
Ingestion points: Commands such as meegle workitem get, meegle view items, and meegle comment list ingest user-generated content (titles, descriptions, comments) from the Feishu Project platform into the agent's context.
Boundary markers: The skill contains a dedicated 'Prompt Injection Defense' section in references/error-handling.md which provides explicit instructions to the AI agent to treat remote data as display text only and never as instructions.
Capability inventory: The skill possesses significant write capabilities, including work item creation/updating, workflow transitions (workflow transition), and attachment management (attachment upload/delete).
Sanitization: Relies on the internal instruction set to ensure the agent maintains context and does not execute commands embedded in retrieved project data.
[EXTERNAL_DOWNLOADS]: The skill documentation provides instructions for installing the necessary CLI utility from an external package registry.
Evidence: npm install -g @tingwillforever/meegle-cli as specified in references/runtime-private-remote-mcp.md.
Context: The package is published under the author's own namespace and is the core functional component required for the skill to operate.

meegle