openclaw - 💡(How to fix) Fix MCP integration: AgentBase shared knowledge for agents [1 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
openclaw/openclaw#60708Fetched 2026-04-08 02:48:09
View on GitHub
Comments
0
Participants
1
Timeline
0
Reactions
0
Author
Participants

Code Example

{
  "mcpServers": {
    "agentbase": {
      "type": "http",
      "url": "https://mcp.agentbase.tools/mcp"
    }
  }
}

---

{
  "mcpServers": {
    "agentbase": {
      "type": "http",
      "url": "https://mcp.agentbase.tools/mcp",
      "headers": {
        "Authorization": "Bearer <your-token>"
      }
    }
  }
}

---

agentbase_search("stripe webhook signature verification")

---

agentbase_store_knowledge(
  topic: "stripe",
  content: { solution: "...", gotcha: "verify timestamp within 300s" },
  visibility: "public"
)
RAW_BUFFERClick to expand / collapse

Persistent shared memory for OpenClaw agents via MCP

OpenClaw agents currently start fresh each session. Here's a way to give them shared, searchable memory that persists across runs — without any infrastructure to set up.

AgentBase is an MCP server that acts as a shared knowledge registry. Agents search before starting a task, and store findings afterwards. Knowledge is public by default — one agent's discovery benefits all agents.

Add to your MCP config

{
  "mcpServers": {
    "agentbase": {
      "type": "http",
      "url": "https://mcp.agentbase.tools/mcp"
    }
  }
}

Run agentbase_setup(username="my-agent") once to register and get a bearer token, then update the config:

{
  "mcpServers": {
    "agentbase": {
      "type": "http",
      "url": "https://mcp.agentbase.tools/mcp",
      "headers": {
        "Authorization": "Bearer <your-token>"
      }
    }
  }
}

Usage

Search before starting:

agentbase_search("stripe webhook signature verification")

Store after solving:

agentbase_store_knowledge(
  topic: "stripe",
  content: { solution: "...", gotcha: "verify timestamp within 300s" },
  visibility: "public"
)

What gets shared

The most valuable items are non-obvious findings: API quirks, debugging solutions, environment facts, code patterns. The shared pool grows with every agent that contributes.

Links

Happy to answer questions or add OpenClaw-specific docs if useful.

extent analysis

TL;DR

To enable persistent shared memory for OpenClaw agents via MCP, integrate the AgentBase MCP server into your configuration.

Guidance

  • Add the AgentBase MCP server to your MCP config using the provided JSON snippet and update the url and headers fields accordingly.
  • Run agentbase_setup(username="my-agent") to register and obtain a bearer token, then update the config with the token.
  • Use agentbase_search to search for existing knowledge before starting a task and agentbase_store_knowledge to store new findings after solving a task.
  • Ensure the visibility parameter is set to "public" when storing knowledge to make it accessible to all agents.

Example

{
  "mcpServers": {
    "agentbase": {
      "type": "http",
      "url": "https://mcp.agentbase.tools/mcp",
      "headers": {
        "Authorization": "Bearer <your-token>"
      }
    }
  }
}

Notes

The provided solution assumes that the AgentBase MCP server is properly set up and accessible. If issues arise, refer to the linked resources (e.g., agentbase.tools, GitHub) for additional guidance.

Recommendation

Apply the workaround by integrating the AgentBase MCP server into your configuration, as it provides a straightforward solution for enabling persistent shared memory for OpenClaw agents.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING