Skip to main content
Glama
Sign In
nock

nock

by wrong-nebula
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Local

Nock

Nock

Nock sits between a coding agent (ChatGPT / Codex, Cursor, Claude Desktop, …) and your local tools.

The cloud model does the thinking. A tiny on-device model (Needle) tries to pick the right tool. Destructive actions wait for a confirm token. If Needle is unsure or the request is off-topic, Nock refuses without sending your tool list to the cloud.

The host only sees four tools: nock.resolve, nock.call, nock.search, nock.status.

Install

python3 -m venv .venv
source .venv/bin/activate
pip install -e ".[dev]"

Optional — so nock.resolve can pick tools instead of always refusing:

pip install -e ".[needle]"

That extra pulls JAX. First resolve also downloads the ~14 MB Needle engine.

Related MCP server: agentic-governance-gateway

Point Codex / ChatGPT desktop at it

They share ~/.codex/config.toml. Easiest:

codex mcp add nock -- /ABS/PATH/nock/.venv/bin/nock serve --warmup

Use the absolute path to the venv nock binary. Restart the app, then /mcp should list nock.

Your tools live here, not in Codex

~/.config/nock/config.toml:

[discover]
include_host_configs = false

[servers.fs]
command = "npx"
args = ["-y", "@modelcontextprotocol/server-filesystem", "/ABS/PATH/to/a/folder"]
default_risk = "safe"

[servers.fs.tools.write_file]
risk = "destructive"

Keep Codex’s MCP list to Nock only. Nock starts the filesystem (and anything else you add).

How the host should use it

  1. nock.resolve with the user text.

  2. fast → nock.call the returned envelope.

  3. shortlist → pick one of the listed tools, then nock.call.

  4. escalate → answer without tools. Do not call nock.search.

  5. If nock.call returns type: nock.confirm, call again with confirm=true and confirm_token.

Reads should not confirm. Writes should.

Commands

nock serve          # stdio MCP server (what the host launches)
nock status         # dry config check (does not talk to a running serve)
nock doctor         # binaries, Needle, duplicate servers
nock hosts          # copy-paste snippets for Codex, Claude, Cursor, Continue
nock index          # warm Needle’s tool index

Tests

pytest

MIT. Needle is a separate project from Cactus Compute.

Available Tools

4 tools
nock.callCall a namespaced upstream MCP toolB
Destructive

Proxy tools/call to an upstream server. Destructive tools return a confirm card unless confirm=true and confirm_token matches.

Inputs: name (required), server (required), confirm, arguments, +1 more.

nock.resolveResolve a user utterance to a local toolA
Read-only

Map natural language to one leaf tool call against the local MCP catalog. One hop only. On mode=fast, pass the envelope to nock.call. On mode=shortlist, pick from shortlist. On mode=escalate, do not ask for more tools and do not invent catalog names.

Inputs: text (required), k. Returns: object

nock.searchSearch the local tool catalogA
Read-only

Return up to k MCP Tool objects matching query via the local Nock ranker. Does not execute anything. Disabled by default.

Inputs: query (required), k. Returns: object

nock.statusNock sidecar healthA
Read-only

Mux health, catalog size, last index time, Needle load state. No secrets.

Inputs: no named parameters. Returns: object

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Deploy Server
    Claude Code, inside the Godot 4 editorFlockbay puts an AI agent in the real Godot 4 editor: it writes GDScript, runs your game and fixes what breaks. Free, or bring your own Claude Code or Codex.flockbay.comAd

    TDQS

    A4/5.0

    Scored across 4 tools

    Disambiguation5/5

    Each tool has a clearly distinct role: resolve handles natural-language routing, call executes concrete tool envelopes, search retrieves catalog metadata without executing, and status reports health. No two tools appear to do the same thing.

    Naming Consistency5/5

    All tools share the consistent nock.* prefix and use short, lowercase action-oriented names. The pattern is uniform and predictable across the set.

    Tool Count5/5

    Four tools is well-scoped for a gateway/catalog server: discovery, resolution, execution, and status. Each tool earns its place with no redundancy.

    Completeness5/5

    The tool surface covers the full proxy lifecycle: search the catalog, resolve natural language to a call, execute via upstream, and check server health. No obvious missing operation for the stated purpose.

    Maintenance

    ActivityInactive
    ResponsivenessOutdated data

    Resources

    Related MCP Connectors

    Related MCP Servers

    • projscan

      abhiyoheswaran1
      A
      license
      Not graded
      quality
      C
      maintenance
      An MCP server that provides AI coding agents with AST-accurate, context-budget-aware codebase querying, safety gates, and team policy integration via structured tools and a local plugin layer.
      112 npm
      4
      MIT

    • worker-mcp

      noosxe
      A
      license
      A
      quality
      B
      maintenance
      MCP server that enables a coordinator AI agent to spawn, control, and supervise local coding agents with interactive gating for high-risk operations.
      10
      10 npm
      MIT

    • LeanRigor MCP Server

      badrabbitt
      A
      license
      Not graded
      quality
      C
      maintenance
      A local-first MCP gateway that reduces context overhead in coding-agent sessions by exposing only relevant tools and enforcing task-appropriate engineering gates, with privacy-first defaults and no network calls until explicitly enabled.
      Apache 2.0