Skip to main content

pr-council-mcp

CI Coverage PyPI Python License

Platform support: macOS only for now. Why Linux support is pending.

Illustrative terminal replay of the PR Council review lifecycle

Illustrative, time-compressed replay with a fictional PR. Open the interactive version and transcript.

pr-council-mcp is a playground for agentic engineering end to end. It explores production concerns such as durable workflows, model coordination, sandboxed tools, secrets, human approval, and observability in a local environment with a small footprint: one stdio MCP server backed by SQLite, without a service stack to deploy.

The concrete application is multi-model review of GitHub pull requests. Independent quality and security reviewers inspect a PR, deliberate over their findings, and combine retained issues into a preview. The calling agent shows that preview to a human; only after explicit approval does the server publish inline comments and a COMMENT-only summary to GitHub.

Status: early development (alpha). The design, tool surface, and configuration are still changing and may break between versions. Expect rough edges, and expect the interfaces to evolve as both tracks are explored.

Quickstart

The server currently runs on macOS and requires Python 3.12+. Install uv, git, rg, and gh, then authenticate GitHub:

gh auth login
gh auth status

On first launch, the server creates ~/.config/localmcp/localmcp.toml (or the path beneath $XDG_CONFIG_HOME) with this base configuration if the file does not exist. The llm, secrets, and observability tables are owned by localmcplib; pr-council owns only its optional pr_review table:

schema_version = 1

[llm]
backend = "native"

[secrets.openai_api_key]
env_vars = ["OPENAI_API_KEY"]

[secrets.anthropic_api_key]
env_vars = ["ANTHROPIC_API_KEY"]

On macOS, store the keys in your default Keychain. Each command prompts for the key without putting it in shell history:

security add-generic-password -U -s localmcp -a openai_api_key -w
security add-generic-password -U -s localmcp -a anthropic_api_key -w

Only the providers used by your selected models need keys. macOS may ask you to allow the server to read these Keychain items. The server starts without keys and reports missing credentials when a review is requested.

Install the package on demand from public PyPI in your MCP client configuration.

Claude Code

Add .mcp.json:

{
  "mcpServers": {
    "pr-council-mcp": {
      "command": "uvx",
      "args": ["--from", "pr-council-mcp", "pr-council-mcp"]
    }
  }
}

opencode

Add opencode.jsonc:

{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "pr-council-mcp": {
      "type": "local",
      "command": ["uvx", "--from", "pr-council-mcp", "pr-council-mcp"]
    }
  }
}

How it works

Reviews are checkpointed operations rather than one long MCP request. The client starts a review, polls it, presents the exact revision-bound preview, and publishes only after explicit approval. Restarts can recover unfinished work, and publication revalidates the PR revision and approved payload before writing.

By default the server prepares an isolated managed worktree. Reviewer and deliberator models receive one bounded Bash tool inside a read-only macOS Seatbelt sandbox. They can inspect source and Git history with ordinary local tools, but cannot write files, access the network or credentials, or execute repository code. A local-checkout mode is also available when cloning a large repository is impractical.

The MCP surface intentionally stays small: start, status, preview, commit, and cancel. Published reviews contain inline comments plus a COMMENT-only summary; the server never approves or rejects a pull request.

Tools

Inspect the MCP server's tool schema for complete arguments and return models.

Tool What it does Important behavior and requirements
pr_council_start Starts a durable initial or follow-up PR review and returns an opaque operation ID. Accepts a GitHub PR URL, source mode, finalized context, review mode, optional baseline, iterations, and optional model overrides. It never publishes comments.
pr_council_get Polls review status, progress, failures, and completion data. The calling agent retains the operation ID and polls until the review is ready, terminal, or needs action.
pr_council_preview Returns the exact revision-bound summary and inline comments. Read-only. The calling agent presents this payload for human review without changing it.
pr_council_commit Publishes one approved preview as inline comments plus a COMMENT-only summary. Requires the preview revision and payload hash, revalidates the PR endpoints and authenticated GitHub user, and never approves or rejects the PR.
pr_council_cancel Requests cooperative cancellation of one owned operation. Cancellation is unavailable once publication has begun.

Basic configuration

Configuration lives at ~/.config/localmcp/localmcp.toml, or beneath $XDG_CONFIG_HOME when set. Shared root values are inherited by every local MCP server and [server.pr-council-mcp] overrides this server's application settings. Unknown application keys are rejected. The generated configuration uses the default reviewer matrix.

PR URLs are accepted from github.com by default. To allow a GitHub Enterprise host, configure:

[server.pr-council-mcp.pr_review]
allowed_hosts = ["github.com", "github.enterprise.example"]

Authenticate gh for the additional host with gh auth login --hostname github.enterprise.example. For headless authentication to a GitHub Enterprise Server host, set GH_ENTERPRISE_TOKEN or GITHUB_ENTERPRISE_TOKEN instead; GH_TOKEN and GITHUB_TOKEN apply to github.com and ghe.com hosts.

Secrets resolve from declared environment aliases first and the shared localmcp OS-keyring service second. The native backend resolves openai_api_key and anthropic_api_key as required by the selected models. GitHub authentication is owned by gh, using the host-specific credentials described above. Secret values are never logged.

Structured logs are written only to ~/.local/state/localmcp/pr-council-mcp/logs/pr-council-mcp.log (or beneath $XDG_STATE_HOME) because stdout and stderr carry the MCP protocol. Langfuse tracing is optional and enabled with LOCALMCP_LANGFUSE_ENABLED=true. Tool inputs and outputs remain suppressed unless LOCALMCP_LANGFUSE_CAPTURE_PAYLOADS=true is also set.

License

Apache-2.0

Release files for pr-council-mcp 0.2.1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for pr-council-mcp 0.2.1
File Size Uploaded
pr_council_mcp-0.2.1.tar.gz 52.2 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for pr-council-mcp 0.2.1
File Interpreter ABI Platform
pr_council_mcp-0.2.1-py3-none-any.whl Python 3 none any Details

Total release size: 115.4 kB

Release files / pr_council_mcp-0.2.1.tar.gz

Download URL pr_council_mcp-0.2.1.tar.gz
Size 52.2 kB
Tags Source
SHA-256 checksum
How to use checksums
f5f92c1bca706c962659d46a748c2b8b28e3c0c0a2e7ca81992c4d8dcd6754b0
BLAKE2b-256 checksum
How to use checksums
1ae83ea282618ac398329b66f8e8a1ac6f7db31bab0f3af964aafbe5694bb085
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 28, 2026.

Transparency log

Release files / pr_council_mcp-0.2.1-py3-none-any.whl

Download URL pr_council_mcp-0.2.1-py3-none-any.whl
Size 63.1 kB
Tags Python 3
SHA-256 checksum
How to use checksums
ad8a27cb362a69ebb29f06e3016fcd42721546e2ec48d645c2498834aced42f9
BLAKE2b-256 checksum
How to use checksums
c4472262d040788ec8b6e5a2243eef59e4d9b3d04f266f1f257c5160a953ac6b
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 28, 2026.

Transparency log

Release history Release notifications | RSS feed

This release

0.2.1 This release

2 release files

0.2.0

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page