phabricator-mcp

phabricator-mcp is an MCP (Model Context Protocol) server that wraps Phabricator's Conduit API, exposing MCP tools for interacting with tasks (Maniphest), code reviews/diffs (Differential), repositories and file contents (Diffusion), users, projects, pastes, wiki (Phriction), blogs (Phame), transactions, uploads/files, builds (Harbormaster), owners, activity feed, chat threads (Conpherence), audits, and PHID lookup—via stdio MCP transport.

Evaluated Mar 30, 2026 (21d ago)
Homepage ↗ Repo ↗ DevTools mcp model-context-protocol phabricator conduit-api typescript stdio code-review tasks devtools
⚙ Agent Friendliness
60
/ 100
Can an agent use this?
🔒 Security
56
/ 100
Is it safe for agents?
⚡ Reliability
32
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
80
Documentation
65
Error Messages
0
Auth Simplicity
80
Rate Limits
20

🔒 Security

TLS Enforcement
70
Auth Strength
60
Scope Granularity
30
Dep. Hygiene
45
Secret Handling
70

Uses a Phabricator Conduit API token for authentication (no OAuth). The README recommends allowlisting read-only MCP tools to reduce risk. Rate-limit and error-handling behaviors are not documented here. Dependency and CVE posture cannot be confirmed from provided data.

⚡ Reliability

Uptime/SLA
0
Version Stability
55
Breaking Changes
45
Error Recovery
30
AF Security Reliability

Best When

You want a local/agent-driven integration that can translate natural language into specific, allowlisted Phabricator Conduit actions using MCP tools.

Avoid When

You cannot securely handle long-lived Conduit API tokens or cannot restrict tool permissions to read-only operations for untrusted agent workflows.

Use Cases

  • Search and manage Phabricator tasks (query, create, edit, comment)
  • Perform code review workflows (search revisions, inspect diffs, add inline comments, accept/reject/abandon)
  • Browse repositories and retrieve file contents/history/code search
  • Query builds and build logs for revisions/diffs
  • Look up users/projects/PHIDs and query recent activity
  • Read/write collaboration artifacts like wiki pages, pastes, and blog posts
  • Use an LLM/MCP client to execute Phabricator workflows from natural language

Not For

  • Public-facing or high-risk automation without careful tool permissioning (it can support write operations)
  • Environments that require OAuth-style authentication or federated identity (it uses Conduit API tokens)
  • Use cases needing a web REST/GraphQL API or webhooks from this package (it is primarily an MCP server)

Interface

REST API
No
GraphQL
No
gRPC
No
MCP Server
Yes
SDK
No
Webhooks
No

Authentication

Methods: Phabricator Conduit API token (PHABRICATOR_API_TOKEN)
OAuth: No Scopes: No

Authentication is performed via a Phabricator Conduit API token configured via PHABRICATOR_API_TOKEN or read from ~/.arcrc. MCP-side tool permissions (e.g., Claude allowlist) are used for controlling which MCP tools can be called, but this is not the same as OAuth scopes.

Pricing

Free tier: No
Requires CC: No

No pricing information is provided for the server package itself (MIT-licensed open source). Runtime costs depend on your Phabricator instance and LLM/MCP client usage.

Agent Metadata

Pagination
none
Idempotent
False
Retry Guidance
Not documented

Known Gotchas

  • Write-capable tools exist; clients like Claude may prompt for permissions—use allowlists to restrict to read-only where appropriate
  • Authentication uses a Conduit token; mishandling tokens can grant broad access depending on token permissions
  • Tool naming/permissions must match the MCP client’s allowlist format (example uses mcp__phabricator__<tool_name>)

Alternatives

Full Evaluation Report

Comprehensive deep-dive: security analysis, reliability audit, agent experience review, cost modeling, competitive positioning, and improvement roadmap for phabricator-mcp.

AI-powered analysis · PDF + markdown · Delivered within 30 minutes

$99

Package Brief

Quick verdict, integration guide, cost projections, gotchas with workarounds, and alternatives comparison.

Delivered within 10 minutes

$3

Score Monitoring

Get alerted when this package's AF, security, or reliability scores change significantly. Stay ahead of regressions.

Continuous monitoring

$3/mo

Scores are editorial opinions as of 2026-03-30.

8642
Packages Evaluated
17761
Need Evaluation
586
Need Re-evaluation
Community Powered