Agent Bridge Mcp — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited Agent Bridge Mcp (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Unified memory + agent bridge for Claude Code. One MCP server, 12 tools, zero config.
npx mevoric init --server http://your-server:4000This single command:
~/.claude/.mcp.json.mcp.json files with old entries~/.claude/settings.jsonnewcode-memory, agent-bridge-mcp) if installedRestart VS Code after running.
| Group | Tools |
|---|---|
| Memory | retrieve_memories, store_conversation, judge_memories |
| Bridge | register_agent, list_agents, send_message, read_messages, broadcast |
| Context | share_context, get_context |
| Checkpoints | save_checkpoint, load_checkpoint |
| Event | What It Does |
|---|---|
| SessionStart | Loads shared contexts, pending messages, and checkpoints |
| UserPromptSubmit | Captures prompts + delivers pending agent messages |
| Stop | Pairs prompt+response, POSTs to memory server, saves context + checkpoint |
The memory tools (retrieve_memories, store_conversation, judge_memories) require a backend server that handles semantic search, embedding, and storage. Mevoric is the client — it calls your server's /retrieve, /ingest, and /feedback endpoints.
Without a memory server configured, the 9 bridge/context/checkpoint tools still work fine. Memory tools will return errors gracefully.
Set the server URL during init:
npx mevoric init --server http://your-memory-server:portOr set it as an environment variable:
MEVORIC_SERVER_URL=http://your-memory-server:port| Variable | Default | Description |
|---|---|---|
MEVORIC_SERVER_URL | (none) | Memory server URL (required for memory tools) |
MEVORIC_DATA_DIR | OS-specific app data | Directory for agent data, messages, contexts |
MEVORIC_AGENT_NAME | (auto-generated) | Human-readable name for this agent |
Mevoric runs as a single Node.js MCP server via stdio. Claude Code starts it automatically.
Agent Bridge: Each Claude Code tab registers as an agent with a name. Agents communicate through file-based messaging in the data directory. A heartbeat file tracks liveness — agents that stop writing heartbeats are cleaned up automatically.
Memory: On session end, the Stop hook pairs the user's prompt with the assistant's response and POSTs it to your memory server for storage. On the next session, retrieve_memories searches for relevant past conversations.
Checkpoints: Structured snapshots of your working state (task, files touched, decisions made). Auto-saved on session end, auto-loaded on session start. Expire after 24 hours.
MIT
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.