Historical local qualification: Source 7cf4b26e focused Core 22/22 and Dashboard 26/26 checks passed. The historical bounded local record keeps their exact scope. Historical source 7cf4b26e passed the 3,286/3,286 full checks and all 36 installed families. Independent bounded local evidence acceptance passed; exact-final-head CI is recorded separately from this frozen local proof. Module-specific and broader limits remain below.
The names that identify work as it moves through the Factory: work identities, stored events, command and aggregate prefixes, diagnostics channels, recorded times, codes, protocols and deliberate duplicates. It records what src/ stores and publishes, and the rules new code follows so that every lane names things one way. Nothing here grants Authority or feeds a decision: diagnostics and views built on these names are never Evidence.
Status: The original 7cf4b26e catalogue passed its frozen full execution and independent bounded local evidence acceptance. The neutral protocols retain the separate source/package acceptance scope in their original record; the joint qualification record owns combined-source review and local execution, with exact-final-head CI recorded separately. Neither parent record qualifies the combined source. The historical acceptance record retains its original scope; additive refresh record identifies current source names without changing stored history. The stored-name tables describe current source, including lifecycle scheduling, discovery sources, service access and the bet modules. The D4 candidate now emits Slice WorkRefs and the bind-through-verify delivery diagnostics below. The retained D5 integration and W31 witness/recovery diagnostics are implemented; other trace points, callId and work-trace projection remain planned. Complete D4 Workflow acceptance remains separate from the independently accepted historical 7cf bounded native full-run evidence. A recorded schedule, batch or ask does not prove a live scheduler, a sent message or applied authority.
Source contracts remain in the module references; this page owns only the shared name registry. The catalogue tests compare it with source. Acceptance evidence belongs in the build review.
Work identity
WorkRef
A WorkRef names one unit of work across modules, so records, diagnostics and views can be joined without parsing IDs. It is derived from identities the records already hold. It is never random, never a path, and never an input to an admission or any other decision.
| Form | Unit of work | Derived from | Status |
|---|---|---|---|
slice:<productId>/<sliceId>/<revision> | One frozen Slice of a Product | The gate's productId and sliceRef (<sliceId>/<revision>) | D4 delivery diagnostics form this WorkRef |
run:<state>#<runId> | One fixture Run | <state>, the state digest below, and the Run's runId | The fields are stored; nothing forms the WorkRef |
occurrence:<id> | One lifecycle occurrence | The occurrence ID, such as discovery/<product>/basic/<date> (lifecycle spec §5) | Occurrence IDs exist in lifecycle scheduling; the WorkRef wrapper is planned |
service-use:<productId>/<useId> | One use of an external service | The Product and the service-access use ID | Planned; not in src/ |
- The state digest is the first 12 hex characters of the SHA-256 of the state directory's real path, as UTF-8 bytes. The real path is
realpathSync.nativeof the resolved directory: absolute, with no symbolic link in it. The fixture CLI keeps the journal, the Evidence store, the fences and the workspaces under that path (withStateinlocal-factory.ts). - A relative or linked spelling of one state directory gives one WorkRef. A moved or renamed directory gives its Runs new WorkRefs, and no stored record changes.
- A fixture Run's
runIdis unique only within its state directory, so its WorkRef carries the state digest. - A WorkRef never enters a command ID, an aggregate ID or a payload. New event data carries the fields a WorkRef derives from (for a Slice,
productIdandsliceRef), so a view computes it without parsing an ID. - The fixture WorkRef
run:<state>#<runId>is not Execution's aggregate IDrun:<runId>:#is outside the journal's ID grammar, so the two cannot be confused. - Spans inside a unit of work are identities the records already hold. They are a journal
commandId, a gatedispatchId(32 hex), an Attempt ID (<runId>/<epoch>), an Evidence ref, a prepared intent's digest and a Repository observation ID.
callId
A planned callId identifies one CLI invocation or one HTTP request, for tracing only. It is random per call, for example from crypto.randomUUID(), and appears only in diagnostics messages, the operations log and error bodies (as requestId). It never enters a command ID, payload, event or stored record. The journal replays only an identical command, so a value that differs between two tries of one command would turn its replay into a conflict.
Journal events
A command records its events in its aggregate's history, each with a pending outbox row that nothing delivers (Command journal). An event's type and data are stored history that views read, so a type never changes and a new data shape is a protocol revision.
Naming rule for new events
- A new event type is
<prefix>.<subject>-<past participle>in lower-case kebab words, such asgate.slice-frozenordelivery.verdict-recorded. It fits the journal's ID grammar: 1–128 characters ofA-Z a-z 0-9 . _ : / -, starting with a letter or digit. <prefix>is the module's registered event prefix (Command and aggregate prefixes); no two modules share one.- The subject names what the command recorded, and the past participle what happened to it.
- New event data carries the fields its WorkRef derives from, and
at(Time). - Stored types that predate the rule, such as
run.submitted,attempt.claimedandgate.bound, keep their names, because a rename would rewrite history.
Stored event types
Each table names the emitting function and top-level data fields. verification.reported is retained history only. A record-valued payload follows the linked module contract; the catalogue scanner checks literal fields, not all indirect schemas.
Execution (src/execution.ts), in aggregate run:<runId>:
| Event | Emitted by | Data |
|---|---|---|
run.submitted | submit | runId, pins, budget |
attempt.claimed | claim | runId, attemptId, epoch, worker |
attempt.completed | complete | runId, epoch, evidence |
run.completed | complete, after attempt.completed | runId |
attempt.failed | fail | runId, epoch, reason, evidence |
attempt.interrupted | interrupt; cancel of a running Run | runId, epoch, cleanup |
run.cancel-requested | requestCancel of a running Run | runId, epoch, reason |
run.cancelled | requestCancel of a queued Run; cancelled, after cancel or an interrupt with cancellation pending | runId, reason |
run.exhausted | settle, after a failed or interrupted Attempt that used the Budget's last Attempt | runId, attempts |
Guarded verification (src/local-factory.ts, src/attempt-ledger.ts), in the same fixture journal. The ledger records each event through its private #create, and each record is its aggregate's state.
| Event | Aggregate | Emitted by | Data |
|---|---|---|---|
verification.requested | verification:<runId> | recordRequest, once per Run | The pinned request, which is also the command's payload |
verification.reported | verification:<runId> | Not since 5acdf7c: the Recipe /1 path wrote it (local-factory.ts:397 at 5acdf7c^), and retained /1 journals hold it | attemptId, epoch, result |
attempt.prepared | attempt:<runId>/<token> | recordPreparation | aggregateId |
attempt.reported | report:<runId>/<epoch>/<token> | recordReport | aggregateId |
attempt.fenced | recovery:<runId>/<epoch> | recordFenceProof | aggregateId |
Portfolio (src/portfolio.ts) and Project guidance (src/project-guidance-core.ts):
| Event | Aggregate | Emitted by | Data |
|---|---|---|---|
portfolio.product-registered | portfolio:local | registerProduct, through decideRegistry | product |
portfolio.product-renamed | portfolio:local | renameProduct, through decideRegistry | productId, displayName |
portfolio.observed | portfolio:product:<productId> or portfolio:usage | plan, inside every import (recordObservations, recordUsage, recordUsageObservations) that records, corrects or retracts an observation | observation, digest, supersedes |
guidance.values-set | guidance:product:<productId> | retained v1 journals only; no new emission | productId, values, provenance, digest, recordedAt, revision |
guidance.values-set.v2 | guidance:product:<productId> | setValues, through decideSet; the name is the constant EVENT | productId, values, provenance, digest, recordedAt, revision |
guidance.values-set.v3 | guidance:product:<productId> | setValuesWithRationale | productId, values, provenance, digest, recordedAt, revision, rationale, rationaleDigest, authority, before, after |
guidance.rationale-supplement.v1 | Original pin-addressed supplement aggregate | appendRationaleSupplement | protocol, original, version, digest, recordedAt, attributedTo, authority, recollection, originalRationale, originalChanged |
Product gate (src/product-gate.ts), in aggregate gate:product:<productId>. The names are the EVENT table, and each command emits exactly one event.
| Event | Command, after gate:<productId>: | Emitted by | Data |
|---|---|---|---|
gate.bound | bind | bindStore | binding, boundBy |
gate.granted | grant:<g> | grant | grant |
gate.revoked | revoke:<g> | revoke | grantId, revoked |
gate.stop-requested | stop:<n> | requestStop | stop, stops |
gate.resumed | resume:<n> | clearStop | stops |
gate.owner-taken | own:<token> | takeOwnership | owner, proof |
gate.owner-released | release:<token> | releaseOwnership | id, generation, released |
gate.slice-frozen | freeze:<s>:<r> | freezeSlice | slice |
gate.production-admitted | produce:<hex48> | admitProduction | admission, used, at (since refactor step R4) |
gate.production-recorded | produced:<hex48> | recordProduction | key, outcome |
gate.check-admitted | check:<hex48> | admitCheck | admission, used, at (since refactor step R4) |
gate.dispatch-admitted | dispatch:<id32> | admitDispatch, through #checkDispatch (check) or #integrationDispatch (integrate-local) | dispatch, used |
gate.outcome-recorded | outcome:<id32>:<status or kind> | recordOutcome | id, outcome |
gate.slice-closed | close:<s>:<r> | closeSlice | sliceRef, closed, observationId |
gate.worker-admitted | worker:<dispatchId> | admitWorker | admission, dispatch |
gate.worker-recorded | worker-settled:<dispatchId> | recordWorker | id, outcome |
Delivery records (src/delivery-records.ts): one create-once aggregate per record, written by command record:<aggregate ID> through the private #create. The record itself is the aggregate's state.
| Event | Aggregate | Emitted by | Data |
|---|---|---|---|
delivery.binding-recorded | delivery:binding:<productId> | recordBinding | aggregateId |
delivery.slice-recorded | delivery:slice:<productId>/<sliceId>/<revision> | recordSlice | aggregateId |
delivery.snapshot-recorded | delivery:snapshot:<64 hex> | recordSnapshotManifest | aggregateId |
delivery.verdict-recorded | delivery:verdict:<64 hex> | recordVerdict | aggregateId |
delivery.integration-prepared | delivery:integration:<64 hex> | recordPrepared | aggregateId |
delivery.receipt-recorded | delivery:receipt:<64 hex> | recordReceipt | aggregateId |
delivery.proof-witness-recorded | Original Product/dispatch proof-witness aggregate | recordProofWitness | aggregateId |
Delivery integration requests (src/delivery.ts): create-once original request coordinates, retained before Git preparation or gate admission. The aggregate key hashes Action/occurrence; its command ID record:delivery:integration-request:<64 hex> hashes Product/dispatch ID, binding both immutable identities in one transaction.
| Event | Aggregate | Emitted by | Data |
|---|---|---|---|
delivery.integration-requested | delivery:integration-request:<64 hex> | recordIntegrationRequest | aggregateId |
Delivery source locations (src/delivery-inputs.ts), retained only after freeze:
| Event | Emitted by | Data |
|---|---|---|
delivery.sources-recorded | recordPlanSources | productId, sliceRef, at, sources |
branch-health (src/branch-health-core.ts):
| Event | Emitted by | Data |
|---|---|---|
branch-health.polled | decidePoll | pollId, at, state, changed, snapshotDigest |
branch-health.episodes-opened | decidePoll | productId, pollId, episodes |
branch-health.episodes-failing | decidePoll | productId, pollId, episodes |
branch-health.episodes-closed | decidePoll | productId, pollId, episodes |
branch-health.state-changed | decidePoll | productId, pollId, from, to |
operational-signals (src/operational-signals.ts):
| Event | Emitted by | Data |
|---|---|---|
operational-signals.recorded | #execute | command, receipt |
Milestone objectives (src/milestone-objectives.ts):
| Event | Emitted by | Data |
|---|---|---|
milestone.changed/1 | execute, using the pure transition | protocol, version, commandId, payload, admission, beforeDigest, afterDigest, result |
Research jobs (src/research-jobs.ts), with the event data declared in its pure core:
| Event | Emitted by | Data |
|---|---|---|
research-job.changed/1 | #runFrame | kind, jobId, version, commandId, at, action, data |
Proposal ownership (src/proposal-ownership.ts, src/proposal-ownership-notification.ts):
| Event | Emitted by | Data |
|---|---|---|
proposal-ownership.changed | mutate, using the pure transition | protocol, scope, version, eventId, mutation, admission, result, beforeDigest, afterDigest |
discovery-sources (src/discovery-sources-core.ts):
| Event | Emitted by | Data |
|---|---|---|
discovery-sources.collection-claimed | #claim | occurrenceId, startedAt, deadline |
discovery-sources.collection-started | #register | occurrenceId, day |
discovery-sources.collection-recorded | #record | occurrenceId, digest |
Marketing job (src/marketing-job.ts), in the caller's Journal, aggregate marketing-job:<caller>:<job>:<revision>:
| Event | Emitted by | Data |
|---|---|---|
marketing-job.started | runMarketingJob | job, requestDigest, drafter, budget, asOf, deadline, at |
marketing-job.draft-dispatched | runMarketingJob | job, index, blockId, reservedBytes, at |
marketing-job.draft-settled | settle | job, index, blockId, state, reason, debitedBytes, measured, at |
marketing-job.completed | runMarketingJob | job, packDigest, drafted, at |
lifecycle-schedule (src/lifecycle-schedule.ts):
| Event | Emitted by | Data |
|---|---|---|
occurrence.admitted | applyAdmit | occurrenceId, lane, rhythm, date, week, slotAt, segment, runner, startedAt, deadlineAt, coalesced |
occurrence.resumed | applyResume | occurrenceId, segment, runner, startedAt, deadlineAt |
occurrence.settled | applySettle | occurrenceId, status, reason, occurrence |
schedule.commissioned | recordCommission | revision, digest |
schedule.account-recorded | recordAccount | account, sequence, overage, recordedAt |
occurrence.call-admitted | admitCall | occurrenceId, call, task, account, work, attempt, bound, week, segment |
occurrence.call-ended | endCall | occurrenceId, call, outcome, debit, overBound |
schedule.overage-observed | endCall | account, occurrenceId, call, at |
schedule.sign-in-refused | endCall | account, occurrenceId, call, at |
schedule.window-closed | endCall | account, occurrenceId, call, at, until |
occurrence.deferred | defer | occurrenceId, segment, runner, reason, bound, at, resumeAt |
occurrence.reconciled | reconcile | occurrenceId, segment, unknownCalls, confirmation |
owner-digest (src/owner-digest-core.ts):
| Event | Emitted by | Data |
|---|---|---|
owner-digest.batch-formed | formBatch | slot, slotAt, items, gaps |
owner-digest.digest-formed | formDigest | week, words, readingSeconds, nothingNeedsYou |
owner-digest.note-recorded | recordNote | commandId, week, feeling, outsideAskMinutes |
owner-policy (src/owner-policy.ts):
| Event | Emitted by | Data |
|---|---|---|
owner-policy.channels-recorded | recordChannels | revision, zone, live, channels, recordedBy, recordedAt, version |
owner-policy.ask-recorded | recordAsk | askId, kind, channel, askedAt, subject, productId, effects, minutes, recordedBy, recordedAt, version, answer |
owner-policy.kr1-frozen | recordAsk | askId, askedAt, frozenAt, baseline, evaluation |
owner-policy.ask-answered | answer | askId, answer |
owner-policy.update-recorded | recordUpdate | updateId, kind, channel, sentAt, words, recordedBy, recordedAt, version |
owner-policy.note-recorded | recordNote | week, outsideAskMinutes, feeling, answeredAt, recordedBy, recordedAt, version |
owner-policy, the owner's rules (src/owner-rules.ts):
| Event | Emitted by | Data |
|---|---|---|
owner-policy.rule-proposed | proposeRule | ruleId, class, products, words, validDays, proposedAt, proposedBy, digest, recordedAt, version, confirmations, uses |
owner-policy.rule-confirmed | confirmRule | ruleId, confirmation |
owner-policy.rule-used | recordRuleUse | ruleId, use |
owner-policy.rule-demoted | demote | ruleId, useId, demotion |
prevention-register (src/prevention-register-records.ts), each in its own write-once aggregate:
| Event | Emitted by | Data |
|---|---|---|
prevention-register.result-recorded | recordRegisterResult | key, asOf, productIds, inputDigest, catalogueDigest, digest, counts, at |
prevention-register.exemption-recorded | recordExemption | key, productId, templateId, templateDigest, decidedOn, ownerWords, source, at |
service-access (src/service-access.ts):
| Event | Emitted by | Data |
|---|---|---|
service-access.bound | bindLocation | productId, binding, commissionedBy, atVersion |
service-access.discovered | recordDiscovery | commit, capturedAt, detection, references, findings, tree |
service-access.declared | declareReference | referenceId, location, kind, instance, slot, declaredBy, atVersion |
service-access.confirmed | confirmReference | referenceId, kind, slot, confirmedBy, atVersion |
service-access.route-pinned | declareRoute | kind, instance, purpose, target, declaredBy, atVersion |
service-access.vault-listed | recordVaultListing | vaultId, capturedAt, status, reasons, items, unconventional |
work-items (src/work-items.ts), in aggregates work-items:product:<product> and work-items:index:
| Event | Emitted by | Data |
|---|---|---|
work-items.product-listed | #list | productId, recordedAt |
work-items.product-configured | configure | productId, revision, settings, recordedBy, reason, recordedAt |
work-items.item-proposed | propose | productId, itemId, item, overlaps, recordedBy, reason, recordedAt |
work-items.item-claimed | claim | productId, itemId, item, holder, lane, takeover, overlaps, reason, recordedAt |
work-items.item-advanced | advance | productId, itemId, from, to, links, lane, observedScope, overlaps, actor, reason, recordedAt |
work-items.item-released | release | productId, itemId, holder, closed, reason, recordedAt |
work-items.item-superseded | supersede, conclude | productId, itemId, by, evidence, recordedBy, reason, recordedAt |
work-items.items-evicted | settle | productId, itemIds, comparisonIds, cause, recordedAt |
work-items.comparison-declared | compare | productId, comparisonId, comparison, reason, recordedAt |
work-items.comparison-concluded | conclude | productId, comparisonId, winner, closed, recordedBy, reason, recordedAt |
work-items.lanes-reported | recordReport | productId, observedAt, host, rulesDigest, counts, lanes, items, comparisons, takeovers, collisions, recordedAt |
attempt.* names two families of events on different aggregates with different data: Execution's claims and settlements, and the ledger's records. Both are stored, so both stay.
Implementation worker and Producer records (src/implementation-workflow.ts, src/implementation-dispatch-gate.ts): create-once original records through the module-local retention helper. Data follows the exact retained record contract, not a receipt granting another launch.
| Event | Emitted by | Data |
|---|---|---|
implementation.job-retained | retain | protocol, job |
implementation.request-associated | retain | protocol, origin |
implementation.observation-retained | retain | protocol, jobDigest, dispatchId, evidence, status, cleanup, success, reason |
implementation.proposal-retained | retain | protocol, jobDigest, dispatchId, evidence, digest |
implementation.candidate-retained | retain | protocol, jobDigest, dispatchId, evidence, proposal, candidate, commit, storeDigest |
implementation.settlement-retained | retain | protocol, origin, summary |
implementation.producer-retained | retain | Retained original record: protocol, and the job, request association, worker origin or terminal consumption fields of that exact schema |
Research semantic job (src/research-semantics.ts): the original bounded root/lineage/job/allowance/effect step; no live collaborator is implied.
| Event | Emitted by | Data |
|---|---|---|
research-semantic-step | write | command |
Knowledge (src/knowledge.ts): append-only native source/review/decision/use records under the commissioned current host.
| Event | Emitted by | Data |
|---|---|---|
knowledge.transition.v1 | #write | commandId, request, result, record, stateDigest |
Command and aggregate prefixes
Command IDs and aggregate IDs share the journal's ID grammar. Command IDs are journal-wide: once one is recorded, any other use of it in that journal conflicts (CommandConflictError). So each module's prefixes stay distinct, and no prefix, with its separator, begins another module's.
| Journal file | Holds |
|---|---|
<state>/journal.sqlite, one per fixture state directory | Execution Runs, the fixture request and the Attempt ledger |
<local Portfolio journal> | The Portfolio registry and streams, and Project guidance; the core spec puts the gate and the delivery records there too (§9) |
| Lifecycle, discovery sources, bet modules and the Marketing job | The caller supplies their journal; these libraries do not select a live path |
| Module | Aggregate IDs | Command IDs | Chosen by | Event prefix |
|---|---|---|---|---|
| Execution (reference) | run:<runId> | Any journal ID the caller gives | The caller: the fixture CLI's verify: IDs | run., attempt. |
| Guarded verification (reference) | verification:<runId> | verify:<runId>: then request, submit, claim:<token>, settle:<epoch>, recover:<epoch> or cancel:<uuid>; retained /1 journals also hold report:<epoch> | Derived from the Run, the epoch and the Attempt's owner token; only cancel: takes a fresh random UUID per request | verification. |
| Attempt ledger (reference) | attempt:<runId>/<token>, report:<runId>/<epoch>/<token>, recovery:<runId>/<epoch> | verify:<runId>: then prepare:<token>, report:<epoch>:<token> or fence-proof:<epoch> | Derived | attempt., shared with Execution |
| Portfolio (reference) | portfolio:local, portfolio:product:<productId>, portfolio:usage | Registration and rename: any journal ID; imports: portfolio-import:<64 hex> | The caller; imports derive theirs from aggregate, version and payload | portfolio. |
| Project guidance (reference) | guidance:product:<productId> | New IDs require guidance-; recorded legacy IDs still replay | The caller | guidance. |
| Product gate (reference) | gate:product:<productId> | gate:<productId>: then one of 14 suffixes | Derived from the command's inputs, with the caller's owner token or dispatch ID | gate. |
| Delivery records (reference) | delivery: then binding:, slice:, snapshot:, verdict:, integration:, receipt: or sources:; integration requests use integration-request: | record:<aggregate ID> | Derived | delivery. |
- Shared-journal collisions. Portfolio registration and Execution still accept caller-chosen journal IDs. A collision with a derived command ID can block that command permanently. Guidance now refuses new non-prefixed IDs inside the decision, preserving exact legacy replay.
- Planned service uses. The
service-use:WorkRef is not an implemented use ledger or an authority grant.
Current additional prefixes
| Module | Aggregate IDs | Command IDs | Event prefixes |
|---|---|---|---|
| Lifecycle schedule | discovery/schedule | discovery/commission/<revision>, discovery/account/<account>/<sequence>; occurrence IDs with /admit, /resume/<segment>, /call/<n>, /call/<n>/end, /defer/<segment>, /reconcile/<segment>, /settle | schedule., occurrence. |
| Discovery sources | discovery-sources:collection:<occurrence>, discovery-sources:claim:<occurrence>, discovery-sources:starts:<product>:<day> | discovery-sources/claim/<occurrence>, discovery-sources/record/<occurrence>, discovery-sources/start/<product>/<day>/<version> | discovery-sources. |
| Service access | service-access:bindings, service-access:product:<product> | service-access:<product>:<verb>:<id>; discovery and route use their next sequence, a vault listing its vault id and capture time | service-access. |
| Proposal ownership | proposal-ownership:<64 hex hash of canonical scope> | sfpo:<64 hex hash of scope and original mutation ID> | proposal-ownership. |
| Branch health | branch-health:product:<product> | branch-health/poll/<product>/<poll instant> | branch-health. |
| Owner policy | owner-policy:ledger, owner-policy:rules | owner-policy/channels/<revision>, owner-policy/ask/<id>, its /answer, owner-policy/update/<id>, owner-policy/note/<week>; owner-policy/rule/<ruleId>, its /confirm/<round>, /use/<useId> and /use/<useId>/demote | owner-policy. |
| Marketing job | marketing-job:<caller>:<job>:<revision> | marketing-job/<caller>/<job>/<revision>/ then start, block/<index>/dispatch, block/<index>/settle or complete | marketing-job. |
| Owner digest | owner-digest:main | owner-digest/batch/<slot>, owner-digest/digest/<week>, owner-digest/note/<caller id> | owner-digest. |
| Prevention register | prevention-register:result:<64 hex>, prevention-register:exemption:<64 hex> | prevention-register/result/<64 hex>, prevention-register/exemption/<64 hex>; the hex is the SHA-256 of the record's identity (a result's asOf and Products, an exemption's Product, template and digest) | prevention-register. |
| Work items | work-items:product:<product>, work-items:index; work-items:portfolio is reserved for stage 2 | work-items/index/<product>, work-items/configure/<product>/<revision>, work-items/<verb>/<product>/<24 hex of the item or comparison and the request ID>, work-items/report/<product>/<24 hex of the instant and report digest> | work-items. |
Rule for new prefixes
- A module registers its aggregate prefixes, command-ID prefixes and event prefix on this page in the change that adds its first command. This rule and the event rule apply to the lifecycle and service-access modules from their first increment.
- A command ID is derived from the command's own inputs where it can be, so a retry repeats it exactly. A random part (an owner token, a dispatch ID, a cancellation request) is chosen once and reused for every try of that command.
- A prefix ends with a separator (
:,/or-), and neither it nor any registered prefix begins the other. - The WorkRef forms
slice:,occurrence:andservice-use:are not journal prefixes.
Diagnostics channels
Modules publish diagnostics on node:diagnostics_channel. A diagnostic is never Evidence (core spec G4) and never read by a decision, and the publishing module stores none. publish runs each subscriber synchronously inside the call, which is how fault tests stop, kill or revoke at an exact point.
Rule for new channels and events
- One channel per module, named
sf:<module>. - Event names are
<subject>-<past participle>in lower-case kebab words; the channel already names the module. - A message holds names, identifiers, codes, counts and digests only: never payload text, file content, paths or credentials.
A message is one frozen object with these members.
| Member | When it is present |
|---|---|
event | Always |
productId | The event concerns one Product |
workRef | The event concerns one unit of work |
callId | The call has one |
commandId | A journal command admitted the work |
| Span identities | The event has span fields; each event's row lists them |
Existing channels
These names stay as they are: fault and ordering tests subscribe to them, and renaming an event breaks those tests (as Repository records for its own). Their messages predate the rule: several carry paths, launched carries the sandbox profile text, and sf:producer keys its messages by phase.
| Channel | Constant | Key | Events, then the function that publishes them |
|---|---|---|---|
sf:repository | REPOSITORY_CHANNEL (repository.ts) | event | initial-objects-written (createRepository); objects-written and published (Repository.seal); integration-claimed, integration-withheld, integration-submitted and integration-acknowledged (Repository.invokeOnce) |
sf:confined-executor | EXECUTOR_CHANNEL (confined-executor.ts) | event | counted, materialised, launched and collected (ConfinedExecutor.run); admitted (collect, once the control pipe holds the limit records and the preload's admission record) |
sf:producer | PRODUCER_CHANNEL (producer.ts), published through the contract's publish | phase | prepared and admission (#execute, in claude-producer.ts and openai-producer.ts); spawned, stderr and exited (#exchange, in claude-producer.ts); response (#execute, in openai-producer.ts); finished (settleResult) |
sf:implementation | IMPLEMENTATION_CHANNEL (implementation-workflow.ts) | event | job-retained, attempt-claimed, request-associated, dispatch-reserved, worker-observed, evidence-retained, proposal-retained, candidate-sealed, candidate-retained (event, through explicit literal calls in the worker workflow) |
The sf:delivery events
Delivery implements bind-through-verify events and the bounded D5 candidate's integration-prepared, integration-admitted and integration-receipt-retained, plus W31's proof-witness-retained and proof-recovered on DELIVERY_CHANNEL = "sf:delivery". The other integration events below remain planned; the current source does not publish them. Its events are this closed list, published at exactly these points, and fault tests name them in KILL_AT, STOP_AT, REVOKE_AT and STOP_INTENT_AT (core spec §8). A journal command needs no event, because its command-ID prefix already gives fault points before and after it (KILL_BEFORE, KILL_AFTER). An event marks only what a prefix cannot: the point just after an effect outside the journal and before the record that follows it, or a synchronous window before an effect.
These names supersede the core spec's. Its final-read, invoking, guard, prepared and admitted (§5, I7 and I8; §8, rows A2, A3, A5, A6 and R12) become final-read-passed, invocation-started, guard-entered, intent-prepared and integration-admitted. These planned names stay documented for trace work. The current D5 candidate publishes the three implemented integration rows; Repository publishes its own claim/submission/acknowledgement events.
| Event | Point (core spec §5) | Span fields | Name in the core spec |
|---|---|---|---|
store-created | C2: after createRepository returns, before recordBinding | None | None; the crash point after C2 |
fence-held | O1: after holdFence returns for a slot, before any tryFence or takeOwnership | owner | None; row H kills between O1 and O2 |
predecessor-fenced | O2: after tryFence of the recorded owner returns fenced, before takeOwnership | owner, generation | None |
integration-inspected | R2(b): after inspectIntegration returns, before the one permitted outcome command, if any | dispatchId, intent, observationId, invocation | None; row R13's repeated inspections |
snapshot-stored | F8: after one SnapshotStore.put returns, before recordSnapshotManifest | snapshot | None; the crash point mid-F8 |
candidate-sealed | P2: after seal returns a Candidate, before recordProduction (P3) | commandId, candidate | None; the crash point between P1 and P3 |
collection-started | V5: after the re-export before collection, immediately before collect | commandId, dispatchId, epoch | None |
evidence-collected | V5: after collect returns and the Candidate is re-exported, before judgement (V6) | commandId, dispatchId, epoch | None; the crash point between V3 and V8 |
verdict-issued | V6: after judgeRepository returns a Verdict, before recordVerdict (V7) | commandId, dispatchId, proof, result | verdict-issued (row R12) |
proof-witness-retained | After exact native proof-witness retention, before original outcome recording | dispatchId, proof, witness | Implemented W31 retention boundary |
proof-recovered | After original witness/Verdict recovery settles the same dispatch, without new collectors | dispatchId, proof, witness | Implemented W31 historical recovery |
integration-prepared | After the prepared ledger record and immutable request commit, before Git intent preparation | dispatchId, prepared, occurrence | Implemented D5 retention boundary |
integration-receipt-retained | After the sender's dispatched observation commits to the ledger, before gate outcome recording | commandId, dispatchId, prepared | Implemented D5 receipt boundary |
intent-prepared | I4: after prepareIntegration returns, before recordPrepared (I5) | intent, occurrence | prepared (row R12) |
integration-admitted | I6: after the integrate-local dispatch is admitted, before the final read (I7); never on REPLAYED or a refusal | commandId, dispatchId, intent, occurrence | admitted (rows R12, A2 and A5) |
final-read-passed | I7: after the final read passes, synchronously, before invocation-started | commandId, dispatchId | final-read (I7, row A3) |
invocation-started | I8: immediately before invokeOnce | commandId, dispatchId, intent | invoking (I8; the A2, A3 and A6 row) |
guard-entered | I8: on entry to the guard, after Repository's claim and before the guard's own read | commandId, dispatchId, intent, claim | guard (I8; the A2, A3 and A6 row) |
invocation-returned | I8: after invokeOnce returns an observation, before recordReceipt or recordOutcome (I9) | commandId, dispatchId, intent, observationId, invocation | None |
- Every message carries
eventandproductId, andcallIdwhen the call has one. It carriesworkReffor every event except the first three, which concern the Product and no Slice. commandId, where listed, is the gate command that admitted the work: the production admission forcandidate-sealed, and the dispatch for the others.- Planned trace order (only
integration-admittedand the Repository claim/effect events are currently implemented):integration-admitted,final-read-passed,invocation-started,integration-claimed,guard-entered, thenintegration-withheld, orintegration-submittedandintegration-acknowledged, and lastinvocation-returned.
Time
- A new command's payload carries
at: one ISO 8601 UTC instant with milliseconds, such as2026-10-02T09:00:00.000Z(the gate'sTIMESTAMPform), which the caller reads once. Its event repeats it. A re-issued command with another clock value is another payload, so it conflicts and is re-read, never retried (core spec §1). - The journal records no commit time and is not migrated to add one. Where no record holds a time, a view says "time not recorded" and never interpolates one.
| Record | Time it holds | Whose clock |
|---|---|---|
Evidence metadata (sf-evidence/1) | collectedAt | The collector's |
| Repository collector facts, inside the Evidence bytes | startedAt, endedAt and deadline | The collector's, around ConfinedExecutor.run |
| Fixture worker report | durationMs, a duration only | The monotonic clock |
Repository IntegrationObservation | observation.at | The host's, when the observation completed |
| Gate binding | checkout.capturedAt | The caller's |
| Gate Slice entry | frozenAt, and deadline derived from it | The payload's frozenAt |
| Gate dispatch | admitted.at | The payload's now |
| Gate integration outcome read from an observation | at | The observation's |
| Gate production and check admissions | at in the admission's event, repeating the payload's now (since refactor step R4); none in state. Events recorded before R4 carry no at | The payload's now |
| Delivery records | frozenAt in the Slice record; recordedAt in integration and receipt records | The caller's |
| Guidance revision | recordedAt, in state, result and event but not in the payload | The service's, read once before the transaction, so an identical save still replays |
| Lifecycle schedule | Commission/account times, occurrence and segment times; event fields vary as catalogued above | Injected clock or supplied commission |
| Discovery collection and capture | Collection start/end, capture fetch times and claim deadline | Collector clock |
| Proposal ownership | Generation createdAt, first-offer at, closure at and receipt recordedAt | The qualified fence's locked admission observation; none is physical commit time or proof of receiver contact |
| Branch health | Poll at/pollId, observation time and episode times | Poll caller; provider timestamps remain separately attributed |
| Owner policy | recordedAt plus supplied askedAt, answeredAt, sentAt, proposedAt, confirmedAt, at or demotedAt | Ledger or rules clock and attributed caller time, kept distinct |
| Owner digest | Batch slotAt, digest issue time and note recordedAt | Injected clock; calendar slot is distinct from recording time |
| Prevention register | A result's asOf (when its reads were taken) and an exemption's decidedOn (a day); each record's at, in payload, state and event | The caller's; the module reads no clock |
| Work items | recordedAt of every record, deadline of a comparison, observedAt of a lane report | Injected clock, read once inside each decision; the report's instant is the reader's and is kept apart |
| Service access | Discovery capturedAt; other transitions have atVersion only | Discovery input; no commit time invented |
| Portfolio observation | observedAt (null when unknown) and capturedAt | The source's and the collector's |
| Execution Runs and Attempts, ledger records, the fixture request, Portfolio registrations and the other gate records | None | Not recorded |
Codes
Rule for new results
- A new result carries
code, one word of a closed union of lower-case kebab words, anddetail, 1–280 printable ASCII characters. That is the core spec's text rule: longer text is cut with a trailing..., and any other character becomes?. - Stored records discriminate on
kind; operation results discriminate onstatus. - A refusal is a returned result or a thrown error, never a journal record: refusals record nothing.
- A pre-effect seam's refusal reason (the collector's
CollectGate, Repository'sIntegrationGuard) is<code>: <detail>, cut to the guard's 1–256 printable ASCII characters. - The first result in this form is the gate's
currentCheck(product-gate.ts, refactor step R4), whose codes areCurrentCheckCode;seamAnswer(product-gate.ts) gives its refusal to either seam as that reason. admitDispatchanswers an exact replay withreason: "REPLAYED": history, never a refusal and never a permit.- Existing codes keep their spelling, because callers and tests read them. Composition roots map them to nine families through exhaustive
satisfiestables.
| Family | Meaning |
|---|---|
invalid | The request is malformed or breaks a rule it could have known; nothing was done |
not-found | A named record, Run, Product or file does not exist |
stale | The request names a version, generation or Attempt that a later commit replaced; re-read before deciding again |
conflict | The ID or key is already recorded with other content, or the thing already exists |
limit | A bound, allowance or deadline refuses the request |
corrupt | Stored data failed validation; nothing repairs it silently |
unavailable | A store, process, lock or host feature could not be used, and nothing was started |
unknown-effect | An effect may have happened; its outcome is unknown and permits no blind retry |
internal | A defect in the Factory's own code, or a broken invariant |
A composition root's table for the Evidence store's codes shows the shape. It is a sketch, since no such table exists in src/, and each of these codes has one meaning in evidence.ts.
type CodeFamily = "invalid" | "not-found" | "stale" | "conflict" | "limit" | "corrupt" | "unavailable" | "unknown-effect" | "internal";
const EVIDENCE_FAMILY = {
"invalid-input": "invalid",
"invalid-id": "invalid",
missing: "not-found", // "evidence <id> does not exist"
corrupt: "corrupt",
conflict: "conflict", // "already exists and is not identical"
} as const satisfies Record<EvidenceErrorCode, CodeFamily>;
A missing code, an extra key or a code added to EvidenceErrorCode later fails the typecheck, so the table stays exhaustive. Some codes have more than one meaning: Portfolio's CONFLICT, for one, also reports a stream that kept changing during an import. The reviewed table that first needs such a code maps it, not this page.
No family names a refusal by Authority: the gate's REVOKED, STOPPED, NOT_OWNER, OUT_OF_SCOPE and OWNERSHIP_PROOF, the core spec's DeliveryError denied and the Dashboard's FORBIDDEN. The first composition root that maps them decides their family, in its reviewed table.
Existing vocabularies
| Module | Type, file | Codes | Form |
|---|---|---|---|
| Command journal | JournalErrorCode, errors.ts | INVALID_INPUT COMMAND_CONFLICT VERSION_CONFLICT DECISION_FAILED UNSUPPORTED_SCHEMA STORAGE CLOSED NOT_FOUND READ_ONLY | Upper case |
| Execution | ExecutionErrorCode, execution.ts | RUN_NOT_FOUND RUN_EXISTS RUN_TERMINAL ATTEMPT_ACTIVE STALE_ATTEMPT CANCEL_PENDING CORRUPT_RUN | Upper case |
| Evidence | EvidenceErrorCode, evidence.ts | invalid-input invalid-id missing corrupt conflict | Kebab |
| Attempt fence | FenceErrorCode, attempt-fence.ts | invalid-input exists in-use unavailable storage released | Kebab |
| Attempt workspace | WorkspaceErrorCode, attempt-workspace.ts | invalid-input exists unavailable not-held | Kebab |
| Attempt ledger | LedgerError, attempt-ledger.ts | None: a message only | None |
| Fixture CLI | FactoryErrorCode, factory-support.ts | invalid-input conflict not-found | Kebab |
| Confined executor | ExecutorErrorCode, confined-executor.ts | unsupported-host runtime-unavailable | Kebab |
| Repository | RepositoryErrorCode, repository-protocol.ts (re-exported by repository.ts) | invalid-input rejected exists unavailable missing conflict corrupt git-failed | Kebab |
| Repository collector | SnapshotError, repository-collector.ts | missing corrupt conflict unavailable | Kebab |
| Portfolio | PortfolioErrorCode, portfolio.ts | INVALID CONFLICT NOT_FOUND LIMIT UNAVAILABLE CORRUPT | Upper case |
| Project guidance | GuidanceErrorCode, guidance-input.ts | INVALID NOT_FOUND STALE CONFLICT LIMIT CORRUPT MIGRATION_REQUIRED FORBIDDEN UNAVAILABLE | Upper case |
| Capacity planning | CapacityPlanningError, capacity-planning-contract.ts | INVALID LIMIT CONFLICT CORRUPT | Upper case |
| Material timing | MaterialTimingError, material-timing-contract.ts | INVALID LIMIT CONFLICT CORRUPT | Upper case |
| Prevention register | PreventionRegisterErrorCode, prevention-register.ts | INVALID LIMIT CONFLICT CORRUPT | Upper case |
| Marketing workflow | MarketingErrorCode, marketing-workflow-contract.ts | INVALID LIMIT CORRUPT CONFLICT UNAVAILABLE | Upper case |
| Operational signals | OperationalSignalRefusalCode, operational-signals-contract.ts | invalid unauthenticated unauthorised expired unavailable unsupported-retention reconciliation-required pin-mismatch stale-version page-gap identity-conflict command-conflict capacity corrupt storage-unknown | Kebab |
| Proposal ownership | ErrorCode, proposal-ownership-contract.ts | INVALID UNSUPPORTED CONFLICT STALE FORBIDDEN UNAVAILABLE LIMIT CORRUPT STORAGE UNKNOWN_COMMIT | Upper case |
| Shared notification ownership | NotificationFailureCode, proposal-ownership-notification-contract.ts | INVALID UNSUPPORTED FORBIDDEN UNAVAILABLE CONFLICT STALE LIMIT CORRUPT UNKNOWN_COMMIT | Upper case |
| Milestone objectives | MilestoneFailureCode, milestone-objectives-contract.ts | INVALID LIMIT CONFLICT STALE FORBIDDEN UNAVAILABLE CORRUPT UNKNOWN_COMMIT | Upper case |
| Guidance acceptance | FailureCode, guidance-acceptance-contract.ts | INVALID FORBIDDEN UNAVAILABLE CONFLICT STALE LIMIT CORRUPT; possible commit uses Outcome UNKNOWN | Upper case |
| Provider events | ProviderEventErrorCode, provider-events-contract.ts | INVALID LIMIT UNSUPPORTED SOURCE_CONFLICT FORBIDDEN UNAVAILABLE CORRUPT | Upper case |
| Temporal planning | TemporalPlanningError, temporal-planning-contract.ts | INVALID LIMIT CORRUPT | Upper case |
| Milestone owner interface | MilestoneUiCode, dashboard-contract.ts | INVALID LIMIT CONFLICT STALE FORBIDDEN UNAVAILABLE CORRUPT UNKNOWN_COMMIT | Upper case |
| Dashboard | ApiErrorCode, dashboard-contract.ts | BAD_REQUEST FORBIDDEN NOT_FOUND METHOD_NOT_ALLOWED UNSUPPORTED_MEDIA_TYPE PAYLOAD_TOO_LARGE CONFLICT STALE MIGRATION_REQUIRED BUSY JOURNAL_UNAVAILABLE CORRUPT UNAVAILABLE INTERNAL | Upper case |
| Product gate | GateErrorCode, product-gate.ts | 30 codes, listed in its public interface | Upper case |
| Product gate, current check (a result rather than an error) | CurrentCheckCode, product-gate.ts | not-bound dispatch-unknown dispatch-settled slice-closed revoked stopped not-owner deadline | Kebab |
| Delivery composition | DeliveryError, delivery.ts | invalid-input not-found conflict denied unavailable | Kebab |
| Delivery records | LedgerErrorCode, delivery-records.ts | invalid-input conflict corrupt missing | Kebab |
| Producers | ProducerErrorCode, producer.ts | invalid-input invariant | Kebab |
| Intelligence | RefusalCode, intelligence.ts, a result rather than an error | unknown-task blocked not-specified not-measured escalation-undeclared escalation-mismatch pin-mismatch policy-mismatch proposed not-qualified commission-mismatch implicit-retries producer-unknown not-independent no-eligible-entry | Kebab |
| Research jobs | RefusalCode, research-job-contract.ts | INVALID UNSUPPORTED FORBIDDEN STALE CONFLICT EXHAUSTED UNAVAILABLE CORRUPT UNKNOWN_COMMIT | Upper case |
| Research semantic job | ResearchSemanticsError, research-semantics-contract.ts | INVALID LIMIT CONFLICT FORBIDDEN UNAVAILABLE CORRUPT UNKNOWN_COMMIT | Upper case |
| Knowledge | KnowledgeErrorCode, knowledge-values.ts | INVALID LIMIT FORBIDDEN STALE CONFLICT CORRUPT UNAVAILABLE NOT_FOUND | Upper case |
| Factory workflow | FactoryWorkflowError, factory-workflow-input.ts | INVALID LIMIT CONFLICT UNAVAILABLE | Upper case |
| branch-health | BranchHealthErrorCode, branch-health-contract.ts | INVALID CONFLICT CORRUPT LIMIT | Upper case |
| discovery-sources | DiscoveryErrorCode, discovery-sources-core.ts | INVALID CONFLICT CORRUPT CACHE_UNAVAILABLE CACHE_CORRUPT | Upper case |
| lifecycle-schedule | ScheduleErrorCode, schedule-error.ts | INVALID CORRUPT CONFLICT CONTENDED LIMIT NOT_COMMISSIONED REVISION NOT_A_LANE UNKNOWN_OCCURRENCE UNKNOWN_CALL NOT_RUNNING NOT_OWNER CALL_OPEN SETTLED STALE RULE | Upper case |
| owner-digest | DigestErrorCode, owner-digest-core.ts | INVALID CONFLICT CORRUPT LIMIT | Upper case |
| shared calendar | CalendarErrorCode, calendar.ts | INVALID CORRUPT | Upper case |
| owner-policy | AskLedgerErrorCode, owner-policy.ts | INVALID CORRUPT CONFLICT CONTENDED LIMIT NOT_OPENED REVISION UNKNOWN_ASK RULE | Upper case |
| owner-policy, rules | OwnerRulesErrorCode, owner-rules.ts | INVALID CORRUPT CONFLICT CONTENDED LIMIT RULE ROUND UNKNOWN_RULE UNKNOWN_USE | Upper case |
| service-kinds | AccessErrorCode, service-kinds.ts | INVALID CORRUPT STALE CONFLICT LIMIT TAKEN EXISTS OUT_OF_BINDING NOT_FOUND CLASS ROUTE KIND | Upper case |
| work-items | WorkItemsErrorCode, work-items-core.ts | INVALID LIMIT NOT_CONFIGURED NOT_FOUND DUPLICATE OVERLAP LANDED HELD STATE COMPARISON REVISION CONFLICT CONTENDED CORRUPT GIT_FAILED | Upper case |
Assurance, Repository Assurance and the fixture worker throw a plain TypeError for malformed input. A producer's DispatchGate refuses with GateRejection, whose code is any 1–32 characters of A-Z 0-9 _, not a closed union. One meaning has several words: not found is NOT_FOUND, not-found, missing or RUN_NOT_FOUND, and stale is STALE, VERSION_CONFLICT or STALE_ATTEMPT.
Protocols and frozen names
A stored or pinned shape changes through an explicit revision and compatibility contract, never by silently rewriting history (Factory model, invariant 9). A module may refuse an unsupported old revision; the protocol table records those boundaries. Frozen: every event type, command ID, aggregate prefix and protocol identifier on this page, the sf:repository event names, and the wording of Verdicts.
| Module | Protocol identifiers |
|---|---|
| Provider Guidance tools | sf-provider-guidance-tool-result/1; transport result only, no new persistence or acceptance evidence |
| Milestone objectives | sf-milestone-commission/1, sf-milestone-command/1, sf-milestone-aggregate-key/1, sf-milestone-command-key/1, sf-milestone-frame/1, sf-milestone-journal-command/1, sf-milestone-state/1, sf-milestone-event/1, sf-milestone-completion/1, sf-milestone-result/1, m01/1; value digest family sf-milestone-<kind>/1 |
| Guidance acceptance | sf-guidance-acceptance-commission/1, sf-guidance-acceptance-admission-request/1, sf-guidance-acceptance-admission/1, sf-guidance-domain-acceptance/1; no new persistence, only original Guidance domain evidence |
| Evidence / Assurance | sf-evidence/1: the record envelope |
| Local fixture worker, guarded verification | sf-local-fixture-worker/1, sf-guarded-fixture-worker/1 and sf-guarded-fixture-worker/2 (collectors; guarded /2 is compiled); sf-guarded-fixture/1 and sf-guarded-fixture/2 (source and compiled execution manifests); sf-fixture-gate/1 (the fixture gate's control-pipe records and its start callback's symbol); sf-local-fixture-verify/1, sf-local-fixture-verify/2 and sf-local-fixture-verify/3 (Recipe revisions; /1 is read/settled, /2 source, /3 compiled); sf-local-fixture/1, sf-local-fixture/2 and sf-local-fixture/3 (Candidate kinds); local-fixture-scenario/<digest> (scenario revision) |
| Attempt fence and workspace | sf-attempt-fence/1, sf-attempt-workspace/1 |
| Confined executor | sf-confined-node-stateless/1 (profile), sf-confined-runtime/1, sf-confined-snapshot/1, and sf-confined/1 (the preload's control records) |
| Repository | sf-repository/1, sf-candidate/1, sf-integration/1 |
| Repository Assurance and collector | sf-repository-acceptance/1, sf-repository-verify/1, sf-repository-verify/2, sf-confined-environment/1, sf-repository-scenario-collector/1, sf-repository-scenario-collector/2, sf-repository-observation/1, sf-repository-scenario/1/<hex> |
| Product gate and delivery records | sf-product-gate/1, sf-product-binding/1, sf-slice/1, sf-delivery-snapshot/1; D4 source location record sf-delivery-plan-sources/1; D5 request sf-delivery-integration-request/1; additive sf-product-gate/2 and sf-product-binding/2 retain caller-owned source provenance without rewriting legacy records |
| Portfolio | portfolio-registry/1, portfolio-stream/1, usage-receipt/1 |
| Project guidance | product-guidance/1 (retained digests and revision zero), product-guidance-state/1 (retained state), product-guidance/2 (new digests), product-guidance-state/2 (new state) |
| Dashboard | factory-dashboard/3, served over HTTP and never stored; factory-dashboard-assets/1, the verified private browser-asset manifest; factory-milestone-ui/1, the optional owner interface; factory-milestone-plan/1, passive native plan/date detail; sf-milestone-ui-retention/1 and sf-milestone-ui-originals/1, caller-commissioned private original envelopes; sf-milestone-ui-envelope/1, retained envelope content checksums; sf-milestone-ui-binding/1 and sf-milestone-ui-original-key/1, deterministic binding and original identity domains; factory-dashboard/4 adds explicit caller presentation sf-dashboard-presentation/1; factory-dashboard-assets/2 declares the exact supported pair [factory-dashboard/3, factory-dashboard/4], while manifest /1 supports Dashboard /3 only |
| Capacity planning | sf-capacity-inputs/1, sf-capacity-policy/1, sf-capacity-forecast/1, sf-capacity-outcome/1, sf-capacity-comparison/1; observed-envelope-successive-release/1, a conditional scenario method |
| Material timing | sf-material-timing-inputs/1, sf-material-timing-policy/1, sf-material-timing-result/1 |
| Prevention register | sf-prevention-template/1, sf-prevention-register-input/1, sf-prevention-register-result/1; dispatcher-circuit-breaker/1, the breaker marker a workflow file declares; sf-prevention-register-record/1, the stored record of a result or an exemption |
| Marketing workflow | sf-marketing-fact/1, sf-marketing-brief/1, sf-marketing-pack/1, sf-marketing-validity/1; pure fact-linked packs, never publication. The job adds sf-marketing-job/1 (request and start payload), sf-marketing-job-state/1 (the stored record in the caller's Journal), sf-marketing-job-result/1, and the drafter port's sf-marketing-draft-request/1 and sf-marketing-draft/1 |
| Operational signals | sf-operational-source-binding/1, sf-operational-observation/1, sf-operational-page/1, sf-operational-admission/1, sf-operational-signal/1, sf-operational-intake/1; accepted-source-observation/1, unassessed source facts |
| Proposal ownership | branch-health-notification/1, guidance-set-values/2, sf-first-offer/1, sf-proposal-command/1, sf-proposal-generation/1, sf-proposal-operation/1, sf-proposal-ownership-admission-request/1, sf-proposal-ownership-admission/1, sf-proposal-ownership-commission/1, sf-proposal-ownership-event/1, sf-proposal-ownership-mutation/1, sf-proposal-ownership-page/1, sf-proposal-ownership-receipt/1, sf-proposal-ownership-state/1; closed ownership values and inert target profiles, never receiver acceptance |
| Shared notification ownership | notification-transport/1, sf-proposal-notification-delivery-commission/1, sf-proposal-notification-read/2, sf-proposal-notification-transition/2, sf-proposal-original-mode/1, sf-proposal-ownership-event/2, sf-proposal-ownership-notification-commission/2, sf-proposal-ownership-notification-marker/2, sf-proposal-ownership-notification-mutation/2, sf-proposal-ownership-notification-observation/2, sf-proposal-ownership-notification-receipt/2, sf-proposal-ownership-state/2, sf-provider-event-attempt-marker/1, sf-provider-event-attempt-request/1, sf-provider-event-attempt/1, sf-provider-event-body/1, sf-provider-event-contact-admission/1, sf-provider-event-contact-terminal/1, sf-provider-event-delivery-view/1; fixed original aggregate mode, inert transport observations and canonical body; no sender |
| Provider subscriptions | sf-mcp-subscriptions-private/1, sf-mcp-subscriptions-policy/1; caller-owned private lifecycle state and fixed policy |
| Provider Events RPC | sf-provider-events-rpc/1; inert decoded-method profile, no stored events or diagnostics |
| Provider events | branch-health-polled/1, sf-branch-health-polled-source/1, sf-provider-event-detail/1, sf-provider-event-id/1, sf-provider-event-notification/1, sf-provider-event-permission/1, sf-provider-event-profile/1; scoped historical notification only |
| Temporal planning | sf-temporal-source/1, sf-date-claim/1, sf-temporal-constraint/1, sf-temporal-policy/1, sf-temporal-inputs/1, sf-temporal-plan/1; same-zone-no-instant-resolution/1, the fixed civil-date policy |
| Provider diagnostics | sf-provider-diagnostic-input/1, sf-provider-diagnostic/1; caller-supplied retained-fact projection, no stored event or authority |
| Research jobs | native-basic/1, sf-citation/1, sf-research-brief/1, sf-research-evidence-cursor/1, sf-research-evidence-page/1, sf-research-job-command/1, sf-research-job-event/1, sf-research-job-ref/1, sf-research-job-request/1, sf-research-job-state/1, sf-research-job-view/1, sf-research-job/1, sf-research-locked-admission/1; fixed original collection, attribution and bounded passive reads |
| Milestone work brief | sf-milestone-selected-generation/1, sf-milestone-work-brief-input/1, sf-milestone-work-brief/1, native-factory/1; captured literal snapshot and content identity only, no event, command, admission or authority |
| Milestone Research handoff | sf-milestone-research-handoff/1, recorded-draft/1, sf-milestone-research-link/1, sf-milestone-research-handoff-digest/1, sf-milestone-research-link-digest/1, sf-milestone-research-report/1, sf-milestone-research-review/1; native draft/link digests and optional application-owned record declarations, no module event or authority |
| Intelligence | sf-intelligence/1; task protocols sf-producer-prompt/1, sf-review-prompt/1 and sf-review-observation/1 |
| Producers | sf-producer-request/1, sf-producer-receipt/1, sf-producer-dispatch/1, sf-producer-consumption/1, sf-producer-commission/1, sf-producer-prompt/1, sf-claude-transport-policy/1, sf-openai-transport-policy/1, sf-openai-request/1; sf-claude-executable-pin/1 is caller-owned configuration for the audited CLI version, captured by sf-claude-transport-policy/2; neither is live qualification |
| Implementation worker | sf-execution-implementation-subject/1, sf-implementation-job/1, sf-local-execution-manifest/1, sf-worker-origin/1, sf-confined-local-implementation/1, sf-implementation-candidate/1, sf-implementation-evidence/1, sf-implementation-job-record/1, sf-implementation-observation/1, sf-implementation-proposal/1, sf-local-worker-closure/1, sf-worker-request-association/1, sf-worker-settlement/1; current bounded native record/value vocabulary, no live qualification or additional authority |
| Producer original worker association | sf-producer-implementation-job/1, sf-producer-worker-closure/1, sf-producer-request-association/1, sf-producer-terminal/1; current bounded native record/value vocabulary, no live qualification or additional authority |
| W31 original proof witness | sf-delivery-proof-witness/1, sf-external-check-origin/1, sf-product-gate/2; current bounded native record/value vocabulary, no live qualification or additional authority |
| Implementation Evidence | sf-evidence/2; current bounded native record/value vocabulary, no live qualification or additional authority |
| Atomic Guidance rationale and recollection | product-guidance/3, product-guidance-state/3, guidance-set-values/3, guidance-rationale-supplement-record/1, guidance-rationale-supplement/1, guidance-rationale-supplements/1; current bounded native record/value vocabulary, no live qualification or additional authority |
| Knowledge and local planning shadow | sf-knowledge-history/1, sf-knowledge-retrieval/1, sf-knowledge-state/1, sf-learning/1, sf-knowledge-consolidation/1, sf-knowledge-plan-decision/1, sf-knowledge-planning-shadow/1, sf-plan-input-shadow/1; current bounded native record/value vocabulary, no live qualification or additional authority |
| Research semantic job | sf-research-semantic-allowance/1, sf-research-semantic-job/1, sf-research-semantic-result/1, sf-research-synthesis-input/1, sf-research-synthesis-reply/1, sf-research-triage-input/1, sf-research-triage-reply/1, sf-research-semantic-allowance-state/1, sf-research-semantic-effect/1, sf-research-semantic-lineage/1, sf-research-semantic-root/1, sf-research-semantic-state/1; current bounded native record/value vocabulary, no live qualification or additional authority |
| Factory workflow | sf-factory-workflow/1, sf-factory-workflow-result/1; current bounded native record/value vocabulary, no live qualification or additional authority |
| Reserved | sf-factory-inspect/1: the fixture inspect document, which carries no protocol member; nothing emits the identifier |
| Lifecycle schedule | sf-lifecycle-schedule/1, sf-schedule-commission/1; sf-lifecycle-schedule/2 and sf-schedule-commission/2 use caller identity sf-schedule-identity/1; sf-schedule-command/2 is the instance-bound command hash domain |
| Discovery sources | sf-research-brief/1, sf-capture/1, sf-collection/1, sf-collection-claim/1, sf-collection-starts/1, sf-citation/1, sf-drift/1 |
| Service access | sf-service-access/1, sf-service-access-bindings/1, sf-service-catalogue/1, sf-service-map/1, sf-tree-scan/1 |
| Service catalogue route protocols | github-actions/1, app-store-connect/1, apple-signing/1, google-play/1, android-signing/1 (declarations, not proof of an implemented broker) |
| Branch health | sf-branch-health/2; earlier states are refused, never migrated |
| Owner policy | sf-ask-ledger/1, sf-owner-rules/1 |
| Owner digest | sf-owner-digest/2, sf-digest-commission/1; earlier sf-owner-digest revisions are refused; sf-owner-digest-presentation/1 is the separate presentation-aware state profile, with sf-digest-presentation-input/1 inputs, immutable sf-digest-presentation/1 presentation and sf-digest-presentation-command/1 as the instance-bound command hash domain |
| Work items | sf-work-register/1, sf-work-item/1, sf-work-scope/1, sf-work-similarity/1, sf-work-lanes/1; sf-work-summary/1, a read model that is never stored |
- Verdict wording. A Verdict's reasons are free text inside it, and they are protocol. Three checks compare the canonical JSON of a whole Verdict: the fixture
recheckinguarded-verification.ts, the stored-report checkobservationProbleminfactory-support.ts, and the gate's re-derivation in#integrationDispatchinproduct-gate.ts. A reworded reason would therefore make every stored fixture Verdict invalid and every retained repository proofNOT_VERIFIED. A mapping from a judgement to{code, detail}for display is never stored; refactor step R6 plans it (reasonOf). - Pinned closures. The seven files of
GUARDED_FIXTURE_SOURCESinfixture-manifest.tspin source/2Runs; the seven emitted files ofCOMPILED_FIXTURE_RESOURCESpin compiled/3Runs. A change invalidates the affected historical Verdict; stored manifests retain their original protocol and paths. The compiled-resource adoption draft was accepted locally on 1 October 2026. The eight files ofCOLLECTOR_SOURCESinrepository-collector.tspin source repository recipe/1; the eight emitted files ofCOMPILED_COLLECTOR_RESOURCESpin compiled/2. Changed bytes block dispatch under old pins. Historical Verdicts retain their own expectations and can still be rejudged; that does not establish current validity. Both lists holdevidence.ts. Since refactor step R3 the collector's list holds Repository's wire grammar,repository-protocol.ts, in place ofrepository.ts: a change to Repository's Git, sealing or integration code no longer changes the digest, and a change to the grammar does. sf-evidence/1takes no new record kind. A new kind, such as the lifecycle's Citation (sf-citation/1), is a format of its own, stored outside both closures.- The fixture
inspectdocument is/1by default.FactoryResultinfactory-support.ts, plus the CLI'sexitCode, has noprotocolmember. A reader, such as step 1's run-briefing tool (increment 8), reads a document without one assf-factory-inspect/1and refuses anyprotocolit does not know. A later/2names itself in that member.
Vocabulary
Three terms of the delivery core map to the Factory's one language as follows.
| Term in code | What it is | Factory term |
|---|---|---|
| Allowance | The gate's GrantLimits and SliceEntry.used: production runs, check runs, Attempts per check, integrations and elapsed time | The Slice's Budget. Execution's Budget is per Run |
generation | The gate owner's ownership version: 1 for the first owner, and one higher at each later takeOwnership | The Product-scope epoch (Workers). Execution's epoch is per Run and Attempt |
| Check dispatch | A gate dispatch of kind check, epoch 1 to attemptsPerCheck, with Evidence runId ver-<hex48> and attemptId <runId>/<epoch> | An Attempt of the gate's check Run, not an Execution aggregate. The design defines one Run, so whether delivery Runs become Execution Runs is an open owner decision |
Some exported names mean different things in different files.
| Name | Senses |
|---|---|
CheckOutcome | A check's judgement word (assurance.ts); a check dispatch's outcome (product-gate.ts) |
Admission | A producer's dispatch answer (producer.ts); a gate production or check admission (product-gate.ts) |
LedgerError | The Attempt ledger's error (attempt-ledger.ts); the delivery records' error (delivery-records.ts) |
TokenCounts | Portfolio usage tokens (portfolio-inventory.ts); provider tokens (producer.ts) |
Outcome | The fixture CLI's result word (factory-support.ts); a producer's outcome (producer.ts); neither is the domain Outcome |
Decision | A journal decision (journal.ts); a module's model involvement (intelligence.ts) |
Registry | The Portfolio registry (portfolio.ts); the Intelligence registry (intelligence.ts) |
RepositoryIdentity | A Factory-owned store (repository-protocol.ts, re-exported by repository.ts); a registered checkout (portfolio-inventory.ts) |
MAX_TIMEOUT_MS | 60,000 for a fixture process (local-worker.ts); 300,000 for the executor (confined-executor.ts) |
A file that needs both senses imports one under an alias. A rename waits for the file's next reviewed change, because every one of these files belongs to an accepted module.
Deliberate duplicates
Some definitions exist twice on purpose. A copy stays while its file is frozen, pinned by a receipt, or kept out of another module's import graph. Where a test holds a mirror equal, the table names it. The delivery records' own prepared-intent parser is gone (refactor step R3): the records and the gate parse a prepared intent with parsePreparedIntegration in repository-protocol.ts. The gate's and the records' validator kits, grammars and limit mirrors are gone too (refactor step R4): both take them from strict-input.ts, and the gate re-exports that file's GATE_LIMITS and exports RegExp copies of its PRODUCT_ID, SLICE_ID and GRANT_ID grammars, made from their sources, which neither module validates with. Two of the records' rules still differ from Repository's wire grammar: a Slice revision is bounded by GATE_LIMITS.maxRevision (999 999), where repository-protocol.ts takes any safe integer, and the records' HEAD_REF, still used for observations, is wider than its target-ref rule.
| Duplicate | Where | Why it stays | Held equal by |
|---|---|---|---|
The sf-evidence/1 envelope, which gives a record its identity | evidence.ts (private); EVIDENCE_FORMAT in repository-assurance.ts | evidence.ts is in both closures and exports no envelope function | "the envelope digest the judge recomputes is evidence.ts's own record identity" |
Process classification: classify and classifyObservation | classify in assurance.ts (private); classifyObservation in repository-assurance.ts | assurance.ts is receipt-pinned and exports no classifier | "classifyObservation is exactly the fixture classify over every exit, signal, timeout and collector error" |
The snapshot listing digest: snapshotDigest and listingDigest | snapshotDigest in confined-executor.ts; listingDigest in repository-assurance.ts | Repository Assurance imports only SNAPSHOT_PROTOCOL from the executor | "listingDigest equals the executor snapshotDigest on 200 random listings and refuses what the executor refuses" |
| The executor's limits and profile name | confined-executor.ts; repository-assurance.ts | The same import rule | "the module repeats the executor limits it may not import, and imports no I/O" |
| Canonical JSON | canonicalJson in canonical-json.ts; evidence.ts (private) | The fixture closure stays closed: canonical-json.ts is not in it | The fixture closure pins evidence.ts byte for byte |
PRODUCT_ID | portfolio.ts, project-guidance-core.ts, dashboard-server.ts, repository-assurance.ts, and strict-input.ts, whose grammar the gate and the records validate with (refactor step R4; the gate's exported PRODUCT_ID is a RegExp made from it) | The Portfolio receipt pins portfolio.ts; the historical guidance receipt predates extraction into project-guidance-core.ts. The collector closure holds repository-assurance.ts | Identical patterns; no test |
| Repository's wider product ID | createRepository accepts any journal-grammar ID (DOMAIN_ID in repository-protocol.ts) | Repository's receipts pin the grammar, which is also in the collector's closure, and a binding needs the store's product ID to be the Product's, which PRODUCT_ID bounds, so the gate never binds a wider one | Documented, not narrowed |
The local-fixture-scenario/<digest> formula | local-factory.ts (/1); guarded-verification.ts (/2) | Retained journals hold /1 Runs, and only the owner retires that path | No test |
Changing it safely
What must stay true
- A change that adds an event type, command or aggregate prefix, diagnostics channel or event, protocol identifier or error code updates this page in the same change. A stored name never changes: its tables here grow, and no row is rewritten.
- A new event's data fields, and any field added to a stored event's data, are listed here in the same change, because a new
datashape is a protocol revision. - A change to the
sf:deliverylist comes before the delivery code that publishes it, with the fault tests that name it. - The tables cite current files and symbols. A renamed emitting or publishing function, or a moved constant, fails the catalogue test until this page names the new one.
Tests and helpers
Run the catalogue and documentation tests, then rebuild the site. The catalogue checks stored event names and literal data fields, diagnostics names and keys, protocol identifiers, error-code unions and cited symbols. The documentation map classifies this cross-module reference; it has no matching topic guide.
Limits: the source scanner recognises { type, data } drafts in that order, literal or resolved constant/parameter types, and literal diagnostics event or phase names. An encountered unresolved form fails. Indirect payload schemas, command-ID derivation, clock provenance and duplicate meanings require review. Source-text coverage is not a journal migration, a live-data audit or runtime proof. Delivery's bind-through-verify and three D5 retention/admission events are implemented; its remaining trace events are planned. Diagnostic messages are never retained effect proof.
The Portfolio adoption draft preserves the existing portfolio.* Events, aggregate/command IDs, registry/stream/usage protocols and error classes. Public packaging adds no observation source or authority.
Provider Events RPC emits no diagnostics or stored events. Its constructor ProviderEventsRpcError has fixed INVALID_BINDING; dispatch uses local routing tags and the existing subscription WireError projection. See the adapter reference.
Provider Guidance tools returns the structured Result union in provider-guidance-tools-contract.ts: fixed adapter refusals use INVALID, FORBIDDEN or UNAVAILABLE, while actual receiver Outcomes retain their original codes and historical witnesses. Post-invocation receiver-result-unobserved is a transport observation, never a receipt or a scalar error vocabulary. Constructor validation throws fixed TypeError text.
Provider diagnostics exposes the structured Reading union: diagnostic or unavailable with malformed-input, unsupported-protocol, contradictory-facts or overflow. Secondary attribution preserves the native outcome and remote uncertainty; it is never a scalar retry code, Evidence, or a stored receipt.
Research job refusals preserve original collection ownership. UNKNOWN_COMMIT never authorises another contact; semantic review remains required for attributed collections.
Milestone Research uses the structured Refusal result in src/milestone-research-contract.ts, with fixed INVALID, LIMIT or CONFLICT codes. HandoffInspection and PreparationResult preserve unassessed eligibility and unknown forecast; neither is a Journal receipt or a semantic approval. The adapter emits no stored event, diagnostics channel or new command identity. Its prepared request uses the original M01 command grammar.
Milestone work brief uses BriefResult in src/milestone-work-brief-contract.ts: draft, unavailable or the fixed INVALID, LIMIT, CONFLICT refusal vocabulary. The private Code union in src/milestone-work-brief.ts maps validation failures without disclosing input text. It emits no Journal event or diagnostic channel. The SHA256 content digest is not a Slice, command or dispatch ID.
