sync-brain

name	sync-brain
description	Maintain the Sync brain: ingest sources, synthesize canon, render reports, lint the vault, and audit maturity. Use when the user wants to add evidence to Sync, refresh its research, check its audit score, or edit its doctrine. For running actual work through the council (routing, briefs, Codex dispatch, verification gates, measurement), use the `sync` skill instead, not this one.
argument-hint	ingest | import-run | synthesize | report | visuals | lint | next | audit
license	Proprietary

Sync Brain (maintenance surface)

This is the brain-maintenance skill. If you want to run work through Sync, use the sync skill at skills/sync/SKILL.md. That is the operating layer: routing, briefs, the Claude and Codex council, gates, and measurement.

This skill covers the layer underneath: keeping the doctrine and its evidence honest.

Operate the deployed vault first. Treat CODEX.md, wiki/hot.md, and wiki/index.md as vault-root-relative paths, where the vault root is the directory passed to --vault or opened in Obsidian. In this repo, the template vault root is assets/template-brain/ and the demo vault root is examples/sample-vault/.

Canon rules specific to Sync

• Every canon note states what its evidence does not prove. Enforced by tests/test_canon_index.py.
• The canon index must match disk in both directions. Same test. This exists because an index that advertises missing notes is trusted and wrong.
• Every doctrine claim names the control that enforces it, or it is not doctrine, it is an opinion with a citation.

Secretary: use agents/sync-secretary.md for grounded answers, claim review, and vault maintenance. That secretary reads the brain first, cites a vault note and an official URL, and stays advisory and read-only.

Commands

/sync-brain new <client-slug> --owner <name>
/sync-brain ingest --vault <path> --file <source>
/sync-brain import-run --vault <path> --run-dir <sync-run-directory>
/sync-brain synthesize --vault <path>
/sync-brain report --vault <path>
/sync-brain visuals --vault <path>
/sync-brain lint --vault <path>
/sync-brain next --vault <path>

Source checkout equivalent:

sync-brain new <client-slug> --owner <name>
sync-brain ingest --vault <path> --file <source>
sync-brain import-run --vault <path> --run-dir <sync-run-directory>
sync-brain synthesize --vault <path>
sync-brain report --vault <path> --html-only

Required Operating Rules

1. Read <vault>/CODEX.md.
2. Read <vault>/wiki/hot.md.
3. Read <vault>/wiki/index.md.
4. Preserve .raw/ as immutable source material.
5. Never store credentials in the vault.
6. Never make domain-specific claims without dated trustworthy sources.
7. Keep hot, index, overview, and log current.
8. Record research evidence in references/source-ledger.json.
9. Record domain adapter completion in references/adapter-manifest.json.

Script Mapping

• new -> python scripts/scaffold_vault.py
• ingest -> python scripts/ingest_source.py
• import-run -> sync_brain.adapters.sync_run.import_run
• synthesize -> python scripts/synthesize_brain.py
• report -> python scripts/render_brain_report.py
• visuals -> python scripts/generate_vault_visuals.py
• lint -> python scripts/lint_vault.py
• next -> python scripts/guide_next_action.py

Quality Gates

• No capacity-relief claim unless source and destination have distinct declared pool IDs, observed non-secret authentication modes, and current headroom recorded as available
• No price, quota, or benchmark number without a dated source and the plan or harness it applies to
• No gate result reported from model judgment; a verdict requires a command, an exit code, and captured output
• No commit, push, publish, deploy, spend, account change, or production change without explicit owner approval
• No intentional credential persistence; best-effort redaction and sandbox controls do not prove arbitrary output is secret-free
• No claim of measured improvement without naming the baseline and the measurement

Do not call this brain market-ready unless scripts/audit_brain.py --require market-ready passes. A scaffold is not a finished brain.

Research Refresh

monthly for prices, quotas, model rosters, and plan limits; before every release for official harness requirements; benchmark leaderboards rechecked each refresh because several are silently frozen