No description
  • Go 96.9%
  • Shell 1.4%
  • Dockerfile 1%
  • Makefile 0.7%
Find a file
Ole-Morten Duesund df2253398b feat(store): SQLite with embedded migrations (forgejo-mcp-broker-9jh)
Implements internal/store on top of modernc.org/sqlite (pure-Go, no CGO).
Open applies any pending migrations, Close releases the handle, Ping
underpins /healthz.

Migration design:
- Files embedded via embed.FS under migrations/NNNN_name.sql
- schema_migrations table tracks applied versions; re-open is a no-op
- Each migration runs in its own transaction: no partial commits
- loadMigrations takes an fs.FS so tests can inject synthetic migration
  sets to exercise rollback and conflict paths

Connection pragmas (set via DSN so they apply to every pooled conn):
- journal_mode=WAL — better reader/writer concurrency
- foreign_keys=ON — off by default in SQLite, we always want them
- busy_timeout=5000 — absorb brief contention without surfacing SQLITE_BUSY
- synchronous=NORMAL — standard WAL pairing

Phase 1 schema (0001_initial.sql) is minimal: a broker_meta table with a
schema_version row. Real OAuth tables ship in phase 2.

Tests: 90.1% coverage across public API and internal migration runner,
including bad SQL rollback, PK-conflict record-step failure, and scan
errors on malformed schema_migrations rows.

Closes forgejo-mcp-broker-9jh.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 17:22:47 +02:00
.beads feat(store): SQLite with embedded migrations (forgejo-mcp-broker-9jh) 2026-04-24 17:22:47 +02:00
.claude bd init: initialize beads issue tracking 2026-04-24 16:34:50 +02:00
cmd/broker feat: bootstrap Go project layout (forgejo-mcp-broker-n84) 2026-04-24 16:54:27 +02:00
docs docs: initial planning artifacts for fjmcp-broker 2026-04-24 16:21:01 +02:00
internal feat(store): SQLite with embedded migrations (forgejo-mcp-broker-9jh) 2026-04-24 17:22:47 +02:00
.gitignore bd init: initialize beads issue tracking 2026-04-24 16:34:50 +02:00
AGENTS.md bd init: initialize beads issue tracking 2026-04-24 16:34:50 +02:00
CLAUDE.md bd init: initialize beads issue tracking 2026-04-24 16:34:50 +02:00
go.mod feat(store): SQLite with embedded migrations (forgejo-mcp-broker-9jh) 2026-04-24 17:22:47 +02:00
go.sum feat(store): SQLite with embedded migrations (forgejo-mcp-broker-9jh) 2026-04-24 17:22:47 +02:00
LICENSE bd init: initialize beads issue tracking 2026-04-24 16:34:50 +02:00
Makefile feat: bootstrap Go project layout (forgejo-mcp-broker-n84) 2026-04-24 16:54:27 +02:00
README.md bd init: initialize beads issue tracking 2026-04-24 16:34:50 +02:00

forgejo-mcp-broker

OAuth 2.1 authorization server and MCP session broker for forgejo-mcp.

Lets MCP clients such as Claude.ai connect to a Forgejo instance through a single public HTTPS endpoint, with per-user authentication delegated to Forgejo's own OAuth2 provider. The broker handles the OAuth dance, then spawns a dedicated forgejo-mcp --transport stdio subprocess for each authenticated session, scoped to the authenticated user's Forgejo access token.

Status: Planning. No code yet. See docs/design.md for the architecture and docs/plan.md for the phased implementation plan.

How it fits

Claude.ai ──HTTPS──▶ Caddy ──▶ fjmcp-broker ──stdio──▶ forgejo-mcp  ──▶ Forgejo API
                                  (this)              (one per user     (per-user
                                                       session)          token)
  • fjmcp-broker (this project): one long-running process. Handles OAuth discovery, dynamic client registration, the authorization-code flow against Forgejo, session lifecycle, and stdio-to-streamable-HTTP bridging.
  • forgejo-mcp (existing project): used as-is. Spawned per-session with the authenticated user's FORGEJO_ACCESS_TOKEN in the environment.
  • Caddy: terminates TLS for the public hostname and reverse-proxies to the broker.

Why a broker instead of adding OAuth to forgejo-mcp?

Process-level isolation. Each user's Forgejo token lives in exactly one subprocess — the broker never needs to demultiplex tokens inside a single shared client. This keeps forgejo-mcp's sync.Once singleton-client pattern valid and avoids a refactor of every tool handler. Full trade-off in docs/design.md.

Quick map

File What
docs/design.md Architecture, components, token flow, deployment, security
docs/plan.md Seven-phase implementation plan with acceptance criteria

License

MIT © 2026 Ole-Morten Duesund.