Skip to content

README: document MCP server and both agent-integration paths#15

Merged
kelsi-bizer merged 1 commit into
masterfrom
claude/readme-mcp
May 11, 2026
Merged

README: document MCP server and both agent-integration paths#15
kelsi-bizer merged 1 commit into
masterfrom
claude/readme-mcp

Conversation

@kelsi-bizer

Copy link
Copy Markdown
Owner

Summary

The README was written in PR #13 (Logseq cleanup), before PR #14 added packages/mcp-server. As a result, the MCP integration path was invisible to anyone landing on the repo — even though the server ships and works.

What changed

  • Components list — adds packages/mcp-server.
  • Agent integration section — rewritten to cover both paths side by side:
    • Option A — MCP server (Claude Code, Cursor, MCP-Inspector, any MCP-capable client). Single mcpServers JSON entry in the agent's config.
    • Option B — agentskills.io skill (Hermes, OpenClaw, Claude Code). Symlink the skill into the agent's skill directory.

Both target the same BRAIN_DIR and expose the same four tools. Picking the right one depends on what the agent supports.

Why this matters

For external discovery (ClawHub listing, casual visitors, future alpha testers), the README is the front door. Missing the MCP path made it look like BizerBrain only worked with three named agents, when in fact it works with anything that speaks MCP.

Test plan

  • After merge, the rendered README on GitHub shows the two integration options
  • An MCP-capable agent (e.g. Claude Code) configured per the README's JSON snippet can list and read notes from $BRAIN_DIR

Generated by Claude Code

The README was written in PR #13 (cleanup), before PR #14 added
packages/mcp-server. As a result, the MCP integration path was
invisible to anyone landing on the repo, even though the server
ships in the container build context.

Changes:
- Add packages/mcp-server to the Components list.
- Rewrite the Agent integration section to cover both paths:
  - Option A: MCP server (Claude Code, Cursor, MCP-Inspector, any
    MCP-capable client). Single mcpServers JSON entry.
  - Option B: agentskills.io skill (Hermes, OpenClaw, Claude Code).
    Symlink the skill into the agent's skill directory.

Both target the same BRAIN_DIR and expose the same four tools.
@kelsi-bizer kelsi-bizer marked this pull request as ready for review May 11, 2026 13:47
@kelsi-bizer kelsi-bizer merged commit a0b51e1 into master May 11, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants