Makefile MCP

io.github.ai-ronin-systemsPublished by GitHub account ai-ronin-systemsv0.3.0Updated Oct 25, 2026

Expose trusted GNU Make targets as governed MCP tools without a generic shell.

VerifiedSTDIODesktop onlyDeveloper ToolsSecurity & Monitoring

Overview

AI-generated overview

Turns an existing GNU Makefile into typed MCP tools so an assistant can run only approved make targets instead of a generic shell.

What you can do
  • List and inspect the Make targets discovered in a repository's Makefile
  • Run exposed targets as MCP tools, either one typed tool per target or via generic list, describe and run calls
  • Restrict runs to allow-listed context and target pairs with typed, validated inputs in governed mode
  • Require human approval for tasks marked dangerous, and write hash-chained audit records of each run attempt
  • Preview a target with GNU Make dry-run and check configuration with a doctor command
When to use it
Useful when you want an AI assistant to run a repository's existing test, lint, build or deploy targets through Make, without handing it an arbitrary shell. Best for teams that already keep automation in a Makefile and want explicit exposure, approval and audit of what agents may execute.
Requirements
Local process on Linux or macOS (Windows not supported), with Python 3.11+ and GNU Make installed; run via uvx or pipx. No authentication or environment variables are declared. Governed mode needs a .makefile-mcp.yaml policy file, and policy changes require a restart.
Before you install
Makefile recipes are trusted code and are not sandboxed, so a malicious or buggy Makefile is not contained. Exposed targets can still be destructive, and dangerous tasks rely on the MCP client actually asking a human for approval. Audit logs can be rewritten by a same-user process, and preview is a dry-run, not a sandbox. An agent able to edit unpinned Make code or scripts effectively gains a shell.

Installation

In SourceWeft

  1. Open Makefile MCP in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

Makefile MCP

[PyPI] [Python] [CI] [License]

Expose trusted GNU Make targets to AI agents — without exposing a generic shell or replacing Make with another workflow system.

Makefile MCP turns an existing Makefile into typed MCP tools. It is zero-config and run-safe by default, and deny-by-default when governed.

text
Makefile  →  Makefile MCP  →  Codex / Claude / LangChain / Cursor / VS Code / MCP client

Threat model

Design for a compromised agent: any text it reads can carry a prompt injection. Makefile MCP keeps the worst case short. A hijacked agent can run only the targets you listed, with values you typed, and a human approves the risky ones. Read the threat model for the threats, controls, tests that prove them, and limits.

ProtectedNot protected
Agents cannot run arbitrary shell commands, only exposed Make targets.A malicious or buggy Makefile: recipes are trusted code, not sandboxed.
Without config, agents can only run test, lint, check, and fmt.A target the operator chose to expose doing something destructive.
Governed mode runs only allow-listed (context, target) pairs with typed, validated inputs.Host or container privileges granted outside Makefile MCP.
dangerous tasks need explicit human approval through the MCP client; no approval, no run.An MCP client that auto-approves elicitation without a human.
Every governed run attempt, refused, approved, or cancelled, can be written to an audit log, with secret inputs masked.Secrets a recipe prints in ways that differ from the input value.
The approval prompt shows the exact arguments; agent-supplied text cannot reshape it.An agent that can edit unpinned Make code or scripts: editing a recipe is a shell.
With pin, any edit to the Makefile, its includes, or listed scripts stops the server until a human restarts it.Prompt injection in output the agent does read; agent_output: tail or none narrows it.
Audit records are hash-chained; doctor reports edits. A CI test replays a hijacked agent.A same-user process rewriting the whole log and its chain.

Details: threat model, SECURITY.md, and security model.

Quick start

bash
pipx install makefile-mcp        # or: uv tool install makefile-mcpmakefile-mcp serve

Given:

make
test: ## Run tests	pytest -q
lint: ## Run lint	ruff check .

Makefile MCP exposes make_test and make_lint as MCP tools. Other targets are listed but not runnable until you govern them.

bash
makefile-mcp listmakefile-mcp doctormakefile-mcp run testmakefile-mcp run test --preview   # GNU Make --dry-run; not a sandbox

One-shot use: uvx makefile-mcp serve.

Why Makefile MCP

Makefile MCP is intentionally narrow: the repository's Makefile remains the executable source of truth, while Makefile MCP supplies a governed agent-facing capability boundary.

PrincipleWhat it means
Make stays authoritativeNo duplicate workflow DSL, plugin command language, scheduler, database, or second execution engine.
No generic shell capabilityAgents invoke discovered or explicitly governed Make targets, not arbitrary host commands.
Trusted Makefile, untrusted invocationRepository automation is trusted; caller-selected tasks and declared values are authorized, validated, and bounded. GNU Make's own variable/function semantics are treated as an interpreter boundary.
One execution pathCLI, direct MCP tools, generic MCP tools, preview, contexts, and capabilities share the same catalog, authorization, validation, and executor.
Conservative discoveryMakefile MCP statically exposes only Make syntax it can classify confidently; ambiguous constructs are omitted rather than guessed.
Bounded Unix runtimeForeground execution has bounded input/output, timeout/cancellation cleanup, stdin isolation, process-group handling, and physical-context locking.

See Design rationale for the full capability/property map and the alternatives deliberately left out.

Architecture

text
                         Makefile                            |                    conservative catalog                            |              +-------------+-------------+              |                           |            auto                       governed         no config                   .makefile-mcp.yaml              |                           |              +-------------+-------------+                            |                       Application                  +---------+---------+                  |                   |              MCP direct         MCP generic                  +---------+---------+                            |                         Executor                            |                      exact `make -f`

One catalog. One execution path. Thin CLI and MCP adapters.

See Architecture.

Governed mode

Without .makefile-mcp.yaml, Makefile MCP runs in auto mode: every conservatively discovered root target is listed, but only test, lint, check, and fmt can run, with no caller-controlled Make variables. --legacy-auto-run restores the old run-everything behavior (deprecated).

Add .makefile-mcp.yaml to switch to explicit governance:

yaml
schema_version: 1
contexts:  backend:    directory: backend
tasks:  test:    contexts: [root, backend]    risk: safe
  deploy:    risk: dangerous    variables:      ENV:        type: enum        values: [staging, production]        required: true
capabilities:  verify: test
audit:  path: .makefile-mcp/audit.jsonl
pin:  paths: [scripts]       # freeze Make code and scripts the agent must not change

Only enabled (context, target) pairs are callable. dangerous tasks additionally need human approval (defaults.require_approval), and audit writes one JSON line per call. See Governed mode, Configuration, and Contexts and capabilities.

MCP presentations

Makefile MCP supports three MCP presentations over the same callable catalog. Choose the presentation that best fits your requirements.

bash
makefile-mcp serve --tools direct    # default: one typed tool per callable targetmakefile-mcp serve --tools generic   # list_tasks / describe_task / run_taskmakefile-mcp serve --tools both      # both views, same execution core

See MCP presentations and client setup.

Security

Makefile MCP assumes the repository, Makefiles, included Make code, wrapper scripts, and provisioned tools are trusted operator-controlled code. The security boundary is the invocation: an MCP/CLI caller chooses an exposed task/context and, in governed mode, supplies declared values. Makefile MCP authorizes and validates that request; it does not sandbox hostile Makefiles.

Runtime bounds, cancellation/process-group cleanup, stdin isolation, and locking are primarily execution-correctness properties. They limit failure modes around trusted automation rather than turning Make into a sandbox. The built-in MCP server is stdio-only: no HTTP listener, remote-auth layer, TLS, tenancy, or network-service lifecycle is hidden inside the product.

See SECURITY.md for the security policy, reporting process, trust boundary, and limitations. Detailed implementation notes are in docs/security.md.

Intentional scope limits

  • GNU Make on Linux/macOS; Windows is not currently supported.
  • Trusted repository assumption: Makefiles, includes, scripts, and provisioned tools are operator-controlled code.
  • Static discovery is deliberately conservative, not a complete GNU Make parser.
  • Preview is GNU Make dry-run, not sandboxing; Make evaluation can still have side effects.
  • Direct tool inventory is registered at startup; generic catalog calls can observe supported Makefile changes.
  • .makefile-mcp.yaml is startup policy; changes require restart and fail closed until then.
  • Foreground/bounded tasks only; deliberately detached daemons are outside Makefile MCP's process-group containment model.
  • No HTTP server, auth stack, scheduler, job queue, workflow persistence, plugin execution framework, or arbitrary shell tool.
  • Reload behavior:
    • .makefile-mcp.yaml is startup policy. Restart Makefile MCP after adding, removing, or editing it.
    • Makefile discovery refreshes when the conventional Makefile or tracked literal includes change.
    • Generic calls see the refreshed catalog immediately.
    • Direct tool inventory is a startup snapshot: adding or removing direct tools requires a restart. Existing direct tools still re-check the live catalog when invoked.

Requirements

Python 3.11+ · GNU Make · Linux or macOS

Documentation

  1. Threat model — why agents should not get a shell, threats and controls, the policy and audit log, guarantees and limits. Start here.
  2. Architecture — runtime flow, responsibilities, lifecycle, invariants, and dependency rules.
  3. Design rationale — key tradeoffs, full product-property map, and deliberately rejected alternatives.
  4. CLI reference — commands, options, result model, root detection, and exit codes.
  5. Configuration — authoritative .makefile-mcp.yaml fields, defaults, approval, audit log, and precedence.
  6. Static Make discovery — supported syntax, fail-closed behavior, includes, limitations, and cache semantics.
  7. Governed mode — explicit exposure, typed inputs, risk metadata, human approval, and operating model.
  8. MCP presentations — direct, generic, both, schemas, result/error semantics, naming, and lifecycle.
  9. Contexts and capabilities — monorepo scoping, (context, target) authorization, and semantic mappings.
  10. MCP client setup — Codex, Claude Code, LangChain/LangGraph, Cursor, VS Code, and uvx examples.
  11. Security — trust boundary, Make input safety, arbitrary strings, discovery, environment/path controls, and execution bounds.
  12. Development — contributor workflows, code ownership, regression expectations, and Definition of Done.
  13. Deployment — Docker/stdio deployment, production checklist, and provisioning custom tools plus Makefiles.
  14. Releasing — version/tag contract, clean-package smoke test, PyPI Trusted Publishing, provenance, GitHub Releases, and MCP Registry metadata.

Security reporting: SECURITY.md. Contributing: CONTRIBUTING.md.

Source: README.md at commit f08160c

Tools

0
SourceWeft started this package once in an isolated sandbox, without credentials, to list its tools. That failed: the server did not start.

Version history

2
  1. v0.3.0LatestOct 25, 2026
  2. v0.2.2Oct 15, 2026

Score

50% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts