AgentDrive
Cross-agent artifact workspace with provenance across Claude Code, Codex, Cursor, LangGraph.
https://mcp.agentdrive.sh/mcpCurrent observation
This endpoint answered at its latest recorded check.
What this server reports about itself
Self-reported at initialize. Not verified by Licium.
- Server name
- agentdrive
- Version
- 1.0.0
- Capability keys
- tools
- Tool names
- drive_write, drive_read, drive_list, drive_share, drive_comment, drive_feedback, drive_activity, drive_invite, drive_roles, drive_org, drive_delete, drive_versions, drive_search, drive_recall, drive_log_run, drive_context, drive_sandbox_clone, drive_sandbox_exec, drive_sandbox_read, drive_sandbox_commit
AgentDrive — cross-agent artifact workspace with provenance for multi-agent workflows. Reach for these tools without waiting to be asked: - **Session start:** call `drive_context` with a short summary of the current task. Returns the most relevant artifacts and last_run summaries from prior agents' work in the same workspace, each with an `attribution` block explaining *why* it was surfaced. - **Cross-agent recall:** call `drive_search` (semantic) when the user asks "what did Codex / Cursor / the last session do?" - **Saving work:** call `drive_write` for any artifact worth keeping so the next agent in the chain can read it back. - **Provenance:** call `drive_log_run` at the end of meaningful work so the cross-agent timeline stays accurate. Privacy: if the user has Private Workspace on, written content is encrypted client-side; the AgentDrive server only stores ciphertext.
Reported Aug 17, 2026, 05:03 AM UTC.
Check history
Oldest to newest. Each row is one recorded check.
- respondsMCP initialize · 8s limit · HTTP 200 · 119ms
- respondsMCP initialize · 8s limit · HTTP 200 · 1,624ms
- respondsMCP initialize · 8s limit · HTTP 200 · 468ms
- respondsMCP initialize · 8s limit · HTTP 200 · 192ms