Searchable Continuity and Artifact Indexing Guidance¶
This guide gives operators simple naming and linking conventions that keep continuity artifacts — the durable files that carry a project's state across sessions, such as snapshots, handoff packets, and readiness evidence — easy to find and inspect, without standing up a search system. (See the glossary for the precise definition of a continuity artifact, and How Brain Factory works for the bigger picture.)
Why this exists¶
The framework already has strong lifecycle, milestone, readiness, handoff, and queue guidance. Even so, operators lose time reconstructing:
- the latest continuity snapshot
- the latest handoff/resume packet
- current readiness evidence
- blocked/deferred queue posture
- current lifecycle/milestone state
This guide adds lightweight naming/linking/indexing conventions so those answers are discoverable quickly from durable artifacts.
Scope and boundary¶
- This is a docs/template convention layer only.
- Keep indexing lightweight and human-maintainable.
- Do not add a runtime indexing engine or full search service.
Canonical continuity artifact groups¶
| Group | Canonical artifacts | What should be obvious |
|---|---|---|
| Lifecycle + milestone state | runbooks/framework-lifecycle-map.md, framework-state-milestones.md |
Current stage, last milestone, implied next step |
| Continuity snapshots | framework-continuity-snapshot-template.md, runbooks/create-continuity-snapshot.md |
Latest snapshot, previous snapshot, next action |
| Handoff/resume packets | handoff-packet-template.md, runbooks/resume-from-handoff-packet.md |
Latest handoff packet, resume readiness, read order |
| Readiness evidence | runbooks/apply-setup.md, framework-readiness-checklist.md |
Latest readiness status, evidence links, freshness |
| Queue/deferred state | framework-queued-execution-memory.md, runbooks/operate-framework-task-queue.md, runbooks/run-queue-health-check.md |
What is blocked, deferred, pending, or in progress |
Lightweight continuity index pattern¶
For each active workstream, keep one canonical continuity index block in the primary issue or PR. Keep it short and link outward.
Use this structure:
## Continuity artifact index
- Workstream: <short name>
- Current lifecycle stage: <stage 1-8 + link>
- Milestone status reference: <latest milestone log link>
- Read this first: <latest continuity snapshot link>
- Latest continuity snapshot: <link>
- Latest handoff packet: <link or n/a>
- Latest readiness evidence: <link>
- Latest queue/deferred status: <link>
- Blocked by / deferred summary: <one line + links>
- Next ready action: <one line + target artifact>
- Last updated (UTC): <timestamp + owner>
Naming and search-token conventions¶
Use stable heading tokens so repository search can find current artifacts fast. Keep tokens unchanged and update only the values/links.
| Artifact class | Recommended heading token |
|---|---|
| Continuity index | ## Continuity artifact index |
| Snapshot body | ## Continuity snapshot |
| Handoff packet body | ## Handoff packet |
| Resume verification note | ## Resume verification |
| Milestone status block | ## Framework milestone acknowledgments |
| Queue/deferred summary block | ## Queue and deferred-work posture |
| Readiness evidence block | ## Readiness evidence |
Linking and freshness rules¶
- Link latest ↔ previous snapshots. Every new snapshot should include the prior snapshot link and the canonical index link.
- Link handoff packets to snapshots and index. Handoff packets should point to the latest snapshot and canonical continuity index.
- Link readiness evidence from snapshot/index. Avoid leaving readiness proof only in CI logs or chat.
- Link queue/deferred source artifacts directly. Prefer queue file, queue health output notes, and queue-linked issues over summaries without links.
- Update one index per state change. Refresh the continuity index whenever lifecycle stage, readiness, handoff status, or queue posture changes.
Fast-answer map for operators¶
Use the continuity index first, then confirm in linked artifacts.
| Operator question | Read first | Then confirm in |
|---|---|---|
| What state is this workstream in? | Continuity index (Current lifecycle stage) |
Lifecycle map + milestone log |
| What should I read first? | Continuity index (Read this first) |
Latest continuity snapshot |
| What is the latest continuity artifact? | Continuity index (Latest continuity snapshot) |
Snapshot timestamp + previous link chain |
| What is blocked, deferred, or ready next? | Continuity index (Latest queue/deferred status + Next ready action) |
Queue health + queue file + active issue/PR |
| Is resume safe now? | Latest handoff packet + resume verification note | Readiness evidence + milestone status |
Low-maintenance safeguards¶
- Keep one canonical continuity index per active workstream.
- Prefer links over duplicated narrative.
- When information diverges, update the index first, then deeper artifacts.
- If no handoff exists yet, keep
Latest handoff packet: n/aexplicit.