The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of the memi CLI to check integration status, synchronize a documentation corpus, and execute research synthesis tasks.
[EXTERNAL_DOWNLOADS]: Fetches markdown documentation files (.md, .mdx, .markdown, .mdoc) from remote repositories via the memi mermaid-jam corpus sync command to build a local analysis corpus.
[REMOTE_CODE_EXECUTION]: Includes instructions for the agent or user to execute npm install and npm run build within a local checkout directory to prepare the plugin manifest.
[PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external content, creating a surface for indirect prompt injection.
Ingestion points: Processes user-supplied Mermaid source code, local documentation files (e.g., ./docs/flow.md), and markdown files synchronized from remote repositories.
Boundary markers: The instructions do not specify any delimiters or warnings to ignore instructions embedded within the processed diagram source or markdown files.
Capability inventory: The skill has access to execute CLI tools (memi, npm), write files to the .memoire/ directory, and access local file paths.
Sanitization: There is no documentation of sanitization or validation routines to filter malicious instructions from the ingested diagram source.

mermaid-jam