Files
OwnCord/docs
Claude 0dd3e86bc0 feat(b3-7): alpha-shaped dataset — deterministic seed profile and the v1.2.0-alpha.4 snapshot
go run ./cmd/seed -confirm-dev -profile alpha fills an empty database with
the plan's dataset: 100 users (1/2/5/92 across the four roles), 12 channels
(10 text + 2 voice; 3 role-override, 2 user-override, 1 archived), 20,000
messages over 30 simulated days on a diurnal curve (exactly 15% in DMs
across 40 pairs), 300 attachment rows (60/10/10/20%, 10KB-5MB), 500
reactions, 30 invites (10 revoked), one disabled plugin row. Deterministic
by construction — fixed seed, fixed clock, constant bcrypt hash, explicit
ids and timestamps, VACUUM INTO as the canonical bytes — and
TestAlphaProfileByteIdentical holds the property (two full runs compared
byte for byte; a schema_versions wall-clock leak was the one leak found,
now pinned by the scrub). Two constants deliberately leave no rows and say
why in the package comment: voice sessions are LiveKit-ephemeral, and the
replay log is empty exactly as on a server restarted for an upgrade.

The committed snapshot (3.2MB, under the 5MB LFS line) is the scrubbed
VACUUM of that profile at the alpha.4 migration set - the schema has not
moved since the tag, so it is a true alpha.4 artifact. scrub.sql beside it
also anonymises a real donated database. db/alpha_snapshot_test.go is the
standing canary: provenance (31 applied migrations), HEAD migrations apply
cleanly, and every promised row count checks out, FTS included. Consumers
(B4 HP-4, B6 upgrade rehearsal, B10 in-place upgrade) are named in the
snapshot README and docs/deployment.md.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01B8dwVLEihnGZYtH9X631F4
2026-08-31 05:14:13 +00:00
..

Documentation index

Every document in docs/ is listed here. If it is not on this page it is not current guidance.

Docs fall into four kinds, and the difference matters when you are deciding whether to trust one: guidance tells you how to do something, reference describes a contract the code actually implements, audits are dated snapshots that were true when written and were never updated, and plans record intent. Read an audit as history, not as status.

Start here

I want to… Read
Run a server quick-start.md
Deploy for real deployment.md
Contribute a change contributing.md
Understand the system architecture/
Report a bug Issues
Ask, or suggest an idea Discussions
Report a vulnerability security.md
Know who can read what trust-model.md

Guidance

Document Covers
quick-start.md Getting a server running with the fewest steps.
deployment.md Production deployment on Windows and Linux.
contributing.md Environment setup, the branch and PR model, coding standards, how to run the checks CI runs.
security.md How to report a vulnerability, and how findings are handled in public vs private.
trust-model.md Who can read what: operator-readable text and files, E2EE media, transport, at rest, what beta does not claim.
livekit-setup.md Standing up the LiveKit SFU for voice and video.
port-forwarding.md Making a server reachable from outside the LAN.
tailscale.md Remote access without port forwarding.
mcp-introspect.md Dev-only MCP server for introspecting a running instance.

Reference

These describe contracts the code implements. If one disagrees with the code, the code is right and the document is a bug.

Document Covers
api.md REST API under /api/v1.
protocol.md WebSocket protocol — frames, sequencing, reconnect.
schema.md SQLite schema and migrations.
server-configuration.md Every server configuration option.
credential-storage.md What the desktop client persists, and where.
../protocol/schema.json Generated-code source of truth, at the repository root because it is owned by neither side. Server/ws/message_types.go and Client/src/lib/protocolTypes.ts are generated from it — never hand-edit either. See ../protocol/README.md.

Architecture

architecture/README.md indexes the blueprints and carries the maintenance rule: each blueprint names its source-of-truth files, and a PR touching those updates the blueprint in the same change.

client-architecture.md is a redirect stub; the live document is architecture/client.md.

Audits — dated, not maintained

Point-in-time snapshots. They are not updated as the code moves, and they are deliberately left alone when paths change, so links from commit messages keep resolving. Anything here may be stale; the ledger and the plan index carry current status.

Audit Scope
audit-2026-08-23-repository-layout.md Repository layout and contributor experience (RL-01RL-22).
audit-2026-08-23-repository-health.md Full repository health.
audit-2026-08-19.md Repo health. States "0 open findings" — untrue since; see the ledger.
audit-test-coverage-2026-08-19.md Test audit (T-*, a separate register from the OC-* ledger).
audit-2026-08-04-docs-and-coverage.md Documentation accuracy and UI/UX test coverage.
audit-2026-08-04.md Security review.
audit-test-coverage-2026-07-25.md Test-coverage audit.
audit-2026-07-19.md Architecture and spec-conformance review.
audit-2026-04-07.md First comprehensive audit.

Plans

plans/README.md indexes every plan with a recorded state — active, partially implemented, design-only, or shipped — and is the authority over a plan's own header, which can drift.

Where status actually lives

Do not read a defect count, or a "what works" claim, out of a document on this page. Status has owners:

Concern Source of truth
Defect status .superpowers/findings-ledger.json (FINDINGS.md is rendered from it)
Security-sensitive defects Private GitHub Security Advisories
Phase order and gates plans/repo-health-roadmap-2026-08-23.md
Current measured baseline plans/b0-baseline-2026-08-25.md
Generated-code contracts CLAUDE.md, "Generated code — never hand-edit"

A CI job checks that documents on this page do not contradict the ledger's counts. Adding a count to a document means adding it to that check's allow-list in scripts/check-doc-counts.mjs.