@sema-agent/sdk 16.3.0 → 17.0.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +69 -2
- package/dist/errors.d.ts +78 -12
- package/dist/errors.d.ts.map +1 -1
- package/dist/errors.js +91 -10
- package/dist/errors.js.map +1 -1
- package/dist/events.d.ts +8 -0
- package/dist/events.d.ts.map +1 -1
- package/dist/events.js.map +1 -1
- package/dist/index.d.ts +16 -4
- package/dist/index.d.ts.map +1 -1
- package/dist/index.js +10 -2
- package/dist/index.js.map +1 -1
- package/dist/resources/approvals.d.ts +34 -30
- package/dist/resources/approvals.d.ts.map +1 -1
- package/dist/resources/approvals.js +30 -0
- package/dist/resources/approvals.js.map +1 -1
- package/dist/resources/memory.d.ts +4 -0
- package/dist/resources/memory.d.ts.map +1 -1
- package/dist/resources/memory.js +4 -0
- package/dist/resources/memory.js.map +1 -1
- package/dist/resources/rules.d.ts +4 -7
- package/dist/resources/rules.d.ts.map +1 -1
- package/dist/resources/rules.js.map +1 -1
- package/dist/resources/session-sync.d.ts.map +1 -1
- package/dist/resources/session-sync.js +2 -3
- package/dist/resources/session-sync.js.map +1 -1
- package/dist/resources/sessions.d.ts +8 -0
- package/dist/resources/sessions.d.ts.map +1 -1
- package/dist/resources/sessions.js +8 -0
- package/dist/resources/sessions.js.map +1 -1
- package/dist/resources/tasks.d.ts +12 -1
- package/dist/resources/tasks.d.ts.map +1 -1
- package/dist/resources/tasks.js +12 -1
- package/dist/resources/tasks.js.map +1 -1
- package/dist/resources/tool-approvals.d.ts +53 -90
- package/dist/resources/tool-approvals.d.ts.map +1 -1
- package/dist/resources/tool-approvals.js +24 -12
- package/dist/resources/tool-approvals.js.map +1 -1
- package/dist/sync.d.ts +7 -7
- package/dist/sync.d.ts.map +1 -1
- package/dist/sync.js +32 -17
- package/dist/sync.js.map +1 -1
- package/dist/types.d.ts +65 -21
- package/dist/types.d.ts.map +1 -1
- package/dist/types.js +23 -0
- package/dist/types.js.map +1 -1
- package/openapi.yaml +205 -138
- package/package.json +1 -1
package/openapi.yaml
CHANGED
|
@@ -869,7 +869,9 @@ paths:
|
|
|
869
869
|
evicts an already-accepted instruction to make room. Opposite handling from not_running: the
|
|
870
870
|
text is still valid — let the run resume (which drains the queue) or reduce concurrent steers,
|
|
871
871
|
then resend.
|
|
872
|
-
errorCode "steering.duplicate_input_id" (server >=7.3.0 / core 5.14.0
|
|
872
|
+
errorCode "steering.duplicate_input_id" (server >=7.3.0 / core 5.14.0; server >= 7.118.0 adds `detail.differs`, a
|
|
873
|
+
non-empty subset of SteerReplayFacet — contains `text` ⇒ fresh key; otherwise the same instruction, replay with the
|
|
874
|
+
original values and do NOT change the key) — the Idempotency-Key you
|
|
873
875
|
sent is already parked with DIFFERENT steering content (same key + same content is an idempotent
|
|
874
876
|
no-op and 202s). The engine refuses rather than silently picking a side. Reissue with a fresh
|
|
875
877
|
key — this is a new instruction, not a retry.
|
|
@@ -1029,7 +1031,7 @@ paths:
|
|
|
1029
1031
|
Use POST /v1/runs/{taskId}/cancel for a run-level stop.
|
|
1030
1032
|
errorCode "steering.not_running" (reused verbatim) — terminal / not accepting, or the stream
|
|
1031
1033
|
settled in the race window ("run just finished on this replica — no in-flight turn to interrupt").
|
|
1032
|
-
errorCode "steering.duplicate_input_id" — this Idempotency-Key was already accepted with DIFFERENT
|
|
1034
|
+
errorCode "steering.duplicate_input_id" (server >= 7.118.0: read `detail.differs` first, see SteerReplayFacet) — this Idempotency-Key was already accepted with DIFFERENT
|
|
1033
1035
|
content or at a DIFFERENT tier (core's replay identity includes the normalized tier: the same key
|
|
1034
1036
|
that once went through the steer door (default tier) then hits this door (now) is refused —
|
|
1035
1037
|
"idempotent success without the cut" would be a dispositive lie). Reissue with a fresh key.
|
|
@@ -2533,6 +2535,16 @@ paths:
|
|
|
2533
2535
|
'409': { $ref: '#/components/responses/Conflict' }
|
|
2534
2536
|
'422': { $ref: '#/components/responses/UnprocessableEntity' }
|
|
2535
2537
|
'501': { $ref: '#/components/responses/NotImplemented' }
|
|
2538
|
+
'503':
|
|
2539
|
+
description: >-
|
|
2540
|
+
server >= 7.118.0 — `state.mcp_reconnect_unsettled`: the time this endpoint waits for the re-dial to settle ran out (the same
|
|
2541
|
+
constant as the `/mcp` panel, 10 s). Read it as "this endpoint did not wait for a result", NOT "failed": the re-dial is NOT
|
|
2542
|
+
cancelled and still settles (success writes the roster onto this run, failure removes it). The body is `error` + `errorCode`
|
|
2543
|
+
only (no `outcome`) and carries NO `Retry-After`; this is a close-then-dial transaction — confirm the server state before
|
|
2544
|
+
re-sending, and do not use a re-send as a probe. SDK → ServiceStateError (never auto-retried: a non-submit POST).
|
|
2545
|
+
content:
|
|
2546
|
+
application/json:
|
|
2547
|
+
schema: { $ref: '#/components/schemas/ErrorResponse' }
|
|
2536
2548
|
|
|
2537
2549
|
/v1/sessions/{sessionId}/background:
|
|
2538
2550
|
parameters:
|
|
@@ -3322,8 +3334,11 @@ paths:
|
|
|
3322
3334
|
`approval_binding_mismatch` (`errorCode`, `field`: boundCallId|boundInputHash|answer — the third
|
|
3323
3335
|
member since core 5.7.0/RB-459: the content-ask `answer` payload was refused) — your binding does
|
|
3324
3336
|
NOT match the current pending action (TOCTOU/swap). The SDK splits these by `errorCode` into
|
|
3325
|
-
ConflictError vs ApprovalBindingMismatchError
|
|
3326
|
-
to
|
|
3337
|
+
ConflictError vs ApprovalBindingMismatchError (the engine's own code for the same refusal, `checkpoint.invalid_outcome`,
|
|
3338
|
+
maps to that class too). binding_mismatch = SAFETY signal: refetch + re-present
|
|
3339
|
+
to the human, NEVER auto-retry. A `decidedBy: "policy"` allow on a person-required card is the OTHER `invalid_outcome`
|
|
3340
|
+
shape — the non-binding seat `hostDecision` — and answers `resume_outcome_invalid` (`field: "hostDecision"`, card pending;
|
|
3341
|
+
see ApprovalDecidedBy). (S-57, server >= 7.54.0: the `field:"boundCallId"` shape is
|
|
3327
3342
|
intercepted in steady state by the `approval_stale` second trigger source below — the server
|
|
3328
3343
|
compares the echoed boundCallId against the current pending pre-CAS and answers 409
|
|
3329
3344
|
`approval_stale` + `currentPending`; this code's boundCallId arm covers only the remaining race
|
|
@@ -3351,7 +3366,8 @@ paths:
|
|
|
3351
3366
|
decide either resumes the parked host or starts that new run (both land on the byte-identical
|
|
3352
3367
|
acceptance shape) — a host session already busy with another run instead answers the ordinary 409
|
|
3353
3368
|
`conflict.session_active_run`.
|
|
3354
|
-
EIGHTH group (server >= 7.75.0 / S-252, same PARKED WORKFLOW-CHILD lane
|
|
3369
|
+
EIGHTH group (server >= 7.75.0 / S-252, same PARKED WORKFLOW-CHILD lane; the 7.75.0–7.117.x reading — server >= 7.118.0
|
|
3370
|
+
narrows the triggers, see the note after this group): `decide.workflow_parks_unreadable`
|
|
3355
3371
|
— the workflow run's record carries no readable `parks[]` list (the ENGINE's own park-ownership
|
|
3356
3372
|
ledger, and the single source this lane joins the caller's checkpoint against). A record written by
|
|
3357
3373
|
an engine older than 7.17.0 does not carry the key at all, or one entry is damaged and the whole list
|
|
@@ -3375,6 +3391,15 @@ paths:
|
|
|
3375
3391
|
there is no mapper branch, class, or type-union member to delete or keep for it — unlike
|
|
3376
3392
|
`decide.workflow_host_not_parked` above, an OPEN decide.* code this SDK never gave a name to degrades
|
|
3377
3393
|
through the base `DecideError` exactly as it always did, on any server version.
|
|
3394
|
+
🔴 server >= 7.118.0 (core 7.44.0 #1319): routing and redemption read ONE source — the holder seat `parkedBy` the engine writes on
|
|
3395
|
+
a `parks[]` entry (the record naming the card at the highest generation, exactly one ⇒ it takes the decision). The EIGHTH group
|
|
3396
|
+
widens and now has two causes under one code (the body carries no cause word; the sentence differs): an entry written by an older
|
|
3397
|
+
engine (cards still parked at the upgrade) ⇒ resume `runId` and decide again; a record that cannot be read (a malformed entry,
|
|
3398
|
+
or `parks` present but not a list) ⇒ resuming does NOT clear it — an operator must repair that record; the body may add
|
|
3399
|
+
`readableHolderRunId` (a clue). A record with NO `parks` key holds nothing: a card only it names answers 409
|
|
3400
|
+
`conflict.resume_context_unavailable` (no `runId`). NINTH group: `decide.workflow_park_forked` (409, body `runIds`) — the same
|
|
3401
|
+
run was resumed more than once and the copies all hold this card; nothing delivered, the card stays PENDING; resume one of
|
|
3402
|
+
`runIds` and decide again (SDK → `DecideWorkflowParkForkedError`).
|
|
3378
3403
|
SIXTH group (A-075.10, TIME-BASED retry-later pair): `resume.usage_window_exhausted` (#449 G1,
|
|
3379
3404
|
core 5.60.1 — the deployment governance window on this run's ledger key is full; NOTHING was
|
|
3380
3405
|
consumed or unpinned, the SAME token with the SAME decision redeems once the window slides) and
|
|
@@ -4232,32 +4257,14 @@ paths:
|
|
|
4232
4257
|
substitution), the command bytes were unreadable, or this ask belongs to one of the seven
|
|
4233
4258
|
families that never get offers — so do NOT diagnose the deployment's rule lane from one
|
|
4234
4259
|
frame or from that one refusal.
|
|
4235
|
-
|
|
4236
|
-
|
|
4237
|
-
enum: [policy]
|
|
4260
|
+
decidedBy:
|
|
4261
|
+
allOf: [{ $ref: '#/components/schemas/ApprovalDecidedBy' }]
|
|
4238
4262
|
description: >-
|
|
4239
|
-
server
|
|
4240
|
-
|
|
4241
|
-
|
|
4242
|
-
|
|
4243
|
-
|
|
4244
|
-
(`POST /v1/approvals/{sessionId}/decide`, ApprovalDecision.settledBy) — see that schema for
|
|
4245
|
-
the full account. `decision:"deny"` ONLY: an allow-family decision carrying this key is a
|
|
4246
|
-
contradiction, refused BEFORE THE DECISION (400 `request.body_shape`, the same sentence as
|
|
4247
|
-
the durable leg) rather than after — the engine's refusal cannot travel back up this HTTP
|
|
4248
|
-
leg, and by then this door's own side effects (the session allow-all grant, a persisted
|
|
4249
|
-
rule, a DECIDED row that replays as a plain approval) have already landed. Because the
|
|
4250
|
-
refusal precedes the decision, the ask is NOT settled: the card is still open and the same
|
|
4251
|
-
`approvalId` can be answered again. The PARKED late-acceptance arm reads the same key.
|
|
4252
|
-
|
|
4253
|
-
SHAPE, not word: a non-string is a 400. A word core does NOT know travels UNTOUCHED — this
|
|
4254
|
-
door still answers 200 (its 200 only ever meant "your answer was delivered") and the ENGINE
|
|
4255
|
-
refuses it fail-closed, settling as `approver_contract`, NEVER as `policy_refused`: a
|
|
4256
|
-
misspelled word must not be rendered as "the deployment's policy refused".
|
|
4257
|
-
|
|
4258
|
-
No capability bit: an older server ignores unknown keys and answers 200, so the refusal is
|
|
4259
|
-
still read as a person's there. Judge by server version, or read back whether
|
|
4260
|
-
`tool_end.gate.settlement.kind` ever shows `policy_refused`.
|
|
4263
|
+
server >= 7.118.0 — WHO made this decision (one key read the same way by all three decision doors; see
|
|
4264
|
+
ApprovalDecidedBy). ABSENT = `person`. This door's own answers: a `policy` allow on a person-required card
|
|
4265
|
+
and a word outside the closed set are both 400 `request.body_shape`, refused BEFORE the decision (the card is
|
|
4266
|
+
still open; answer it again with the same `approvalId`). 🔴 An older server (no `approvalDecidedBy`
|
|
4267
|
+
capability bit) DROPS this key silently and answers 200 — probe the bit first.
|
|
4261
4268
|
note:
|
|
4262
4269
|
type: string
|
|
4263
4270
|
maxLength: 2048
|
|
@@ -4349,7 +4356,7 @@ paths:
|
|
|
4349
4356
|
application/json:
|
|
4350
4357
|
schema: { $ref: '#/components/schemas/AskDecisionAck' }
|
|
4351
4358
|
'400':
|
|
4352
|
-
description: 'Malformed body — `request.invalid_json` (unparseable) or `request.body_shape` (schema reject, e.g. an `actor` carrying anything but `label`, or — on a server < 7.104.0 — a `settledBy` key). Server >= 7.
|
|
4359
|
+
description: 'Malformed body — `request.invalid_json` (unparseable) or `request.body_shape` (schema reject, e.g. an `actor` carrying anything but `label`, or — on a server < 7.104.0 — a `settledBy` key). Server >= 7.118.0 reads `decidedBy` (see AskDecisionBody.decidedBy): `request.field_conflict` (a `policy` allow on a person-required card), `decided_by_not_in_proof` (direct-door worker), `request.field_invalid` (outside the closed set); the retired `settledBy` key is a `request.body_shape`.'
|
|
4353
4360
|
content:
|
|
4354
4361
|
application/json:
|
|
4355
4362
|
schema: { $ref: '#/components/schemas/ErrorResponse' }
|
|
@@ -6849,16 +6856,6 @@ components:
|
|
|
6849
6856
|
name: { type: string }
|
|
6850
6857
|
description: { type: string }
|
|
6851
6858
|
content: { type: string }
|
|
6852
|
-
manifest:
|
|
6853
|
-
type: object
|
|
6854
|
-
additionalProperties: true
|
|
6855
|
-
deprecated: true
|
|
6856
|
-
description: >
|
|
6857
|
-
RETIRED — server >= 7.116.0 rejects the WHOLE skill when this key is present on a delegated agent's skill (unknown key:
|
|
6858
|
-
`agent.skills must be …`; a top-level skill answers 400 `request.field_invalid` "skill.manifest was retired in core 7.41.0 — …").
|
|
6859
|
-
core 7.41.0 retired the skill manifest (hard break, no alias). Rewrite `manifest.allowTools` to the skill's
|
|
6860
|
-
`allowedTools` (an allow-rule source, not a fence); `manifest.allowPaths` has no replacement (use the session rules'
|
|
6861
|
-
`allowDirs`). Older servers (<= 7.115.x) still accept it; the key stays declared here until the next major.
|
|
6862
6859
|
allowedTools: { type: array, items: { type: string }, description: 'server >= 7.116.0 — the skill''s allow-rule source (see `SkillSpec.allowedTools`); <= 7.115.x rejects the key (unknown key).' }
|
|
6863
6860
|
disallowedTools: { type: array, items: { type: string }, description: 'server >= 7.116.0 — the skill''s deny-rule source (see `SkillSpec.disallowedTools`); <= 7.115.x rejects the key (unknown key).' }
|
|
6864
6861
|
files:
|
|
@@ -7422,6 +7419,14 @@ components:
|
|
|
7422
7419
|
items: { $ref: '#/components/schemas/TaskAgentDefinition' }
|
|
7423
7420
|
description: Per-task sub-agent definitions (roster additions scoped to this task only).
|
|
7424
7421
|
interactiveTools: { type: boolean, description: Mount interactive (prompting) tools. }
|
|
7422
|
+
approvalReach:
|
|
7423
|
+
allOf: [{ $ref: '#/components/schemas/ApprovalReach' }]
|
|
7424
|
+
description: >-
|
|
7425
|
+
server >= 7.118.0 (core 7.43.0 #1354) — declares that NOBODY on this run can answer approvals. Passed verbatim to the
|
|
7426
|
+
engine. The live approver callback is no longer this run's approver (`wiring_manifest` `ask.form: "declared_none"`); on a
|
|
7427
|
+
deployment with a parking facility the first gated call PARKS the run (`paused`) — for "refuse and keep going" use
|
|
7428
|
+
`permissionMode: "dontAsk"`. It never approves or releases anything. A value outside the closed set ⇒ 400
|
|
7429
|
+
`request.field_invalid`. Probe `capabilities.approvalReach` first.
|
|
7425
7430
|
oneShot:
|
|
7426
7431
|
type: boolean
|
|
7427
7432
|
description: >
|
|
@@ -8926,40 +8931,14 @@ components:
|
|
|
8926
8931
|
not inside the HMAC payload, so accepting it unsigned would be tamperable); 501 without an exemption
|
|
8927
8932
|
store. On the parked-background-agent leg the grant lands on the child's ROOT/HOST session (the whole
|
|
8928
8933
|
session tree shares it). The 200 body echoes `rememberApplied` when a store is wired.
|
|
8929
|
-
|
|
8930
|
-
|
|
8931
|
-
enum: [policy]
|
|
8934
|
+
decidedBy:
|
|
8935
|
+
allOf: [{ $ref: '#/components/schemas/ApprovalDecidedBy' }]
|
|
8932
8936
|
description: >-
|
|
8933
|
-
server
|
|
8934
|
-
|
|
8935
|
-
|
|
8936
|
-
|
|
8937
|
-
|
|
8938
|
-
WHY IT EXISTS: without the word, a policy refusal is byte-identical to "a person pressed No", and the
|
|
8939
|
-
model reads "the user doesn''t want to proceed; STOP and wait for the user to tell you how to proceed"
|
|
8940
|
-
— a sentence about a person, in a run where none is coming. With it the engine mints a POLICY sentence
|
|
8941
|
-
instead, this request''s `reason` travels as the POLICY''s words (not as the user''s guidance), and the
|
|
8942
|
-
settlement is recorded as `policy_refused` (`who.party: "none"`).
|
|
8943
|
-
|
|
8944
|
-
CONSTRAINTS: `decision:"deny"` only — only a person approves an action that EXECUTES, so
|
|
8945
|
-
`{decision:"approve", settledBy:"policy"}` is a contradiction, refused by the SERVER before the
|
|
8946
|
-
decision (400 `request.field_conflict`, the SAME SENTENCE as the live door) with the engine''s own
|
|
8947
|
-
pre-CAS refusal (`checkpoint.invalid_outcome`) behind it as depth; the server does NOT silently drop
|
|
8948
|
-
the key. It is therefore MUTUALLY EXCLUSIVE with `remember` (which is approve-only) — one body
|
|
8949
|
-
cannot carry both. The vocabulary''s owner is still core: the server accepts the KEY and relays the
|
|
8950
|
-
VALUE verbatim, so a word core does not know is refused by the engine, never by a server-side
|
|
8951
|
-
second judge of the vocabulary.
|
|
8952
|
-
REFUSED on a direct-door worker (400 `settled_by_not_in_proof` — not inside the HMAC payload, same
|
|
8953
|
-
rule as `remember` / `updatedInput` / `checkpointToken`); a non-direct-door deployment is unchanged.
|
|
8954
|
-
|
|
8955
|
-
SHAPE, not word: a non-string is 400 `request.field_invalid` (same family as `reason must be a
|
|
8956
|
-
string`). A word core does NOT know travels UNTOUCHED and is refused BY THE ENGINE pre-CAS:
|
|
8957
|
-
409 + `errorCode: "resume_outcome_invalid"` + `field: "hostDecision"` + `retriable: true`, and the
|
|
8958
|
-
card STAYS PENDING (re-decide the same coordinates with a corrected body). NOTE this family no
|
|
8959
|
-
longer answers `approval_binding_mismatch`: from server 7.93.3 the `policy_ask` gate splits
|
|
8960
|
-
`invalid_outcome` by `field`, and the non-binding seats (`hostDecision`, `approver`) answer
|
|
8961
|
-
`resume_outcome_invalid` — "you approved the wrong action" is a sentence about a different fault.
|
|
8962
|
-
The binding seats and the field-absent shape are byte-for-byte unchanged.
|
|
8937
|
+
server >= 7.118.0 — WHO made this decision (one key read the same way by all three decision doors; see
|
|
8938
|
+
ApprovalDecidedBy). ABSENT = `person`. This door's own answers: a `policy` allow on a person-required card is
|
|
8939
|
+
refused by the ENGINE before the CAS (409 `resume_outcome_invalid` + `field: "hostDecision"`, the card stays
|
|
8940
|
+
pending); outside the closed set ⇒ 400 `request.field_invalid`; any value on a direct-door worker ⇒ 400
|
|
8941
|
+
`decided_by_not_in_proof`. An older server (no `approvalDecidedBy` bit) refuses the key (400 `request.body_shape`).
|
|
8963
8942
|
|
|
8964
8943
|
ApprovalRow:
|
|
8965
8944
|
type: object
|
|
@@ -9364,6 +9343,17 @@ components:
|
|
|
9364
9343
|
`noteRecorded` (a live-card-only deployment has no row to write, and a flapping store cannot write
|
|
9365
9344
|
one; both answer `false` while the decision itself still stands). This bit only says "this build
|
|
9366
9345
|
knows the key" — same family as `permissionRulesRevoke`, deliberately gated on no facility.
|
|
9346
|
+
approvalDecidedBy:
|
|
9347
|
+
type: boolean
|
|
9348
|
+
description: >-
|
|
9349
|
+
server >= 7.118.0, hardcoded true (a version bit): the three approval decision doors read `decidedBy` (ApprovalDecidedBy).
|
|
9350
|
+
Send it only when this is present; absent = an older server (the durable doors and `plan_review` 400 it, the live
|
|
9351
|
+
`respond` drops it silently and answers 200).
|
|
9352
|
+
approvalReach:
|
|
9353
|
+
type: boolean
|
|
9354
|
+
description: >-
|
|
9355
|
+
server >= 7.118.0, hardcoded true (same word as the request key): TaskRequest.approvalReach is understood. Absent = an
|
|
9356
|
+
older server — do not send the key.
|
|
9367
9357
|
respondFreeFormRules:
|
|
9368
9358
|
type: boolean
|
|
9369
9359
|
description: >
|
|
@@ -12529,6 +12519,13 @@ components:
|
|
|
12529
12519
|
the vocabulary's owner is still core and it has grown before, so read it the same way as everywhere
|
|
12530
12520
|
else: branch known words, keep a `default` arm, and treat an unknown one as "unknown origin", never
|
|
12531
12521
|
as "no origin". A row minted before the bit existed reads absent (unreported), not "no authority".
|
|
12522
|
+
boundGrantDigest: { type: string, description: 'server >= 7.118.0 — a parked `Skill` card''s grant-binding digest; same value as the live frame''s ToolApprovalFrame.boundGrantDigest.' }
|
|
12523
|
+
boundGrants:
|
|
12524
|
+
allOf: [{ $ref: '#/components/schemas/BoundSkillGrants' }]
|
|
12525
|
+
description: 'server >= 7.118.0 — the value `boundGrantDigest` hashes, published ONLY when it hashes to it (the engine judges; a pair that disagrees reads as absent). DEPLOYMENT TEXT: render with contextual escaping.'
|
|
12526
|
+
classifierNoRuling:
|
|
12527
|
+
allOf: [{ $ref: '#/components/schemas/AutoModeFallbackCause' }]
|
|
12528
|
+
description: 'server >= 7.118.0 — only on a parked `classifier_no_ruling` card: why the auto-mode classifier did not rule.'
|
|
12532
12529
|
objective: { type: ['string', 'null'], description: 'task objective, null if unavailable.' }
|
|
12533
12530
|
input:
|
|
12534
12531
|
description: >
|
|
@@ -12790,7 +12787,11 @@ components:
|
|
|
12790
12787
|
them is not known" narrowed to this; it still does NOT say "never consumed"; see `engine_notice`
|
|
12791
12788
|
`task.user_steer_undrained{cause:"crashed"}`). Converged rows only, one row per (taskId, messageId),
|
|
12792
12789
|
newest 500. From server >=7.111.0 (S-966) the rows include SUBAGENT steers (recorded under the run on the
|
|
12793
|
-
receipt; same row shape
|
|
12790
|
+
receipt; same row shape; `messageId` = the subagent steer receipt's).
|
|
12791
|
+
`kind` = the engine LANE the input was queued on at acceptance (server >= 7.118.0 / core 7.44.0 #1362, closed two words):
|
|
12792
|
+
`steer` (priority `now` / `next` / absent) or `follow_up` (priority `later`, queued for the run's end); the crash notices
|
|
12793
|
+
split the same way (`task.user_steer_undrained` / `task.user_followup_undrained`, the latter with
|
|
12794
|
+
`detail: { followUp, cause: "crashed", taskId, sessionId }`). Read both words; branch on `cause`.
|
|
12794
12795
|
ABSENT = this read has nothing to report (zero rows, or the ledger could not be read —
|
|
12795
12796
|
the authoritative notice is the frame on the run stream); the `?message=` single-entry form never
|
|
12796
12797
|
carries it.
|
|
@@ -12800,7 +12801,7 @@ components:
|
|
|
12800
12801
|
additionalProperties: false
|
|
12801
12802
|
properties:
|
|
12802
12803
|
messageId: { type: string }
|
|
12803
|
-
kind: { type: string, enum: [steer] }
|
|
12804
|
+
kind: { type: string, enum: [steer, follow_up] }
|
|
12804
12805
|
taskId: { type: string, description: The run the input was accepted for (its 200/202 receipt). }
|
|
12805
12806
|
acceptedAt: { type: string, description: ISO timestamp. }
|
|
12806
12807
|
cause: { type: string, enum: [crashed] }
|
|
@@ -13600,6 +13601,20 @@ components:
|
|
|
13600
13601
|
🔴 NOT interchangeable with `taskId`: the workflow-child park lane mints no taskId at all (this
|
|
13601
13602
|
decision never landed on any run, so minting one would be a lie), and the other two decide lanes
|
|
13602
13603
|
mint no runId. Branch by `errorCode`, never fall back from one handle to the other.
|
|
13604
|
+
readableHolderRunId:
|
|
13605
|
+
type: string
|
|
13606
|
+
description: >-
|
|
13607
|
+
server >= 7.118.0, on `decide.workflow_parks_unreadable` (409) ONLY and only on its "the record cannot be read" cause: the
|
|
13608
|
+
readable record that holds the card at the highest generation (absent when there is none or a tie — never null). Where the
|
|
13609
|
+
work can be carried on — a CLUE, not a way to unblock this decision (the damaged record `runId` names keeps it undeliverable
|
|
13610
|
+
until it is repaired). SDK reads it as `DecideWorkflowParksUnreadableError.readableHolderRunId`.
|
|
13611
|
+
runIds:
|
|
13612
|
+
type: array
|
|
13613
|
+
items: { type: string }
|
|
13614
|
+
description: >-
|
|
13615
|
+
server >= 7.118.0, on `decide.workflow_park_forked` (409) ONLY: the resumed copies of one workflow run that all hold this card at
|
|
13616
|
+
the same generation (>= 2). Recovery = resume ONE of them, then decide again; retrying the decide never succeeds.
|
|
13617
|
+
SDK reads it as `DecideWorkflowParkForkedError.runIds`.
|
|
13603
13618
|
cause:
|
|
13604
13619
|
type: string
|
|
13605
13620
|
description: >-
|
|
@@ -14256,11 +14271,14 @@ components:
|
|
|
14256
14271
|
unattended leg run under `UNATTENDED_PERMISSION_MODE=dontAsk`).
|
|
14257
14272
|
AskClass:
|
|
14258
14273
|
type: string
|
|
14259
|
-
enum: [content_question, hook, ancestor, unverifiable_delete, transcript_write, org_rule, governance_unavailable, rule_unreadable, ask_rule, denial_limit_fallback, requires_user_interaction, operator_forced, shell_floor, write_protection_unresolved, probe, write_protection, tool_marks_mcp, tool_marks_a2a, policy_rra, artifact_shared_read, mode, unstated, ordinary]
|
|
14274
|
+
enum: [content_question, hook, ancestor, unverifiable_delete, transcript_write, org_rule, governance_unavailable, rule_unreadable, ask_rule, denial_limit_fallback, classifier_no_ruling, requires_user_interaction, operator_forced, shell_floor, write_protection_unresolved, probe, write_protection, tool_marks_mcp, tool_marks_a2a, policy_rra, artifact_shared_read, mode, unstated, ordinary]
|
|
14260
14275
|
description: >
|
|
14261
|
-
core `ASK_CLASSES` (twenty-two words as of core 7.40.0, twenty-three as of core 7.41.0 — 7.34.0 minted twenty; `shell_floor` joined in 7.39.0 #1260 and rides
|
|
14276
|
+
core `ASK_CLASSES` (twenty-two words as of core 7.40.0, twenty-three as of core 7.41.0, twenty-four as of core 7.43.0 — 7.34.0 minted twenty; `shell_floor` joined in 7.39.0 #1260 and rides
|
|
14262
14277
|
server >= 7.113.0; `unstated` joined in 7.40.0 and rides server >= 7.115.0; `requires_user_interaction` joined in 7.41.0 and rides
|
|
14263
|
-
server >= 7.116.0) — which CLASS an ask falls in (the columns of the mode × class arbitration table).
|
|
14278
|
+
server >= 7.116.0; `classifier_no_ruling` joined in 7.43.0 #1353 and rides server >= 7.118.0) — which CLASS an ask falls in (the columns of the mode × class arbitration table).
|
|
14279
|
+
`classifier_no_ruling` = the auto-mode classifier did not rule (it could not run — `error` / `timeout` — or replied with no verdict —
|
|
14280
|
+
`parse_error`), so a person decides: the denial-limit fallback's row (bypass-immune, no allow rule clears it, the classifier is not asked
|
|
14281
|
+
again); the live frame and the inbox row also carry the cause as `classifierNoRuling`.
|
|
14264
14282
|
`requires_user_interaction` = the MCP server declared, in the tool's advertisement, `_meta["anthropic/requiresUserInteraction"] === true`
|
|
14265
14283
|
(only the boolean `true`): EVERY call of that tool asks the human, in EVERY mode including `bypassPermissions` (it is not one of the ten
|
|
14266
14284
|
RELEASABLE classes), and only the human's answer clears it — `settings.permissions.allow`, operator `APPROVAL_GRANT`, a skill's tool list +
|
|
@@ -15963,18 +15981,12 @@ components:
|
|
|
15963
15981
|
offer on another replica is not a promise that it is still redeemable.
|
|
15964
15982
|
persistedRuleShadowed:
|
|
15965
15983
|
type: string
|
|
15966
|
-
|
|
15967
|
-
|
|
15968
|
-
|
|
15969
|
-
|
|
15970
|
-
|
|
15971
|
-
|
|
15972
|
-
For those older servers (server >= 7.13.0 / core 5.25.0 #144, ADDITIVE, "tool_approval" only): this ask MATCHED one of the
|
|
15973
|
-
caller's persisted allow rules and that rule COULD NOT silence it. The value is the shadowed rule's
|
|
15974
|
-
VERBATIM text (secret-redacted + length-bounded). UNTRUSTED for display, same treatment as `message`.
|
|
15975
|
-
🔴 ABSENT != "you have no rule": present only when a rule matched AND could not clear the ask. 🔴 Deliberately SEPARATE
|
|
15976
|
-
from `governanceForced` and not derivable from it: that key answers "who imposed the gate", this one answers "what
|
|
15977
|
-
happened to YOUR rule". The durable dual on those servers is `gate.riskDescriptor.shadowedRule` on the parked row.
|
|
15984
|
+
description: >-
|
|
15985
|
+
server >= 7.13.0 (core 5.25.0 #144), ADDITIVE, "tool_approval" only — this ask MATCHED one of the caller's persisted allow
|
|
15986
|
+
rules and that rule COULD NOT silence it; the value is the rule's VERBATIM text (secret-redacted + length-bounded; UNTRUSTED
|
|
15987
|
+
for display, like `message`). Absent != "you have no rule"; deliberately separate from `governanceForced`.
|
|
15988
|
+
RETIRED: server >= 7.117.0 (core 7.42.0, BREAKING rename, no alias) never mints this key on any face; its successor is
|
|
15989
|
+
`allowRuleShadowed` (all five allow-layer members, not just the rule store). Kept for servers in [7.13.0, 7.117.0) only (sdk 16.3.0).
|
|
15978
15990
|
allowRuleShadowed:
|
|
15979
15991
|
type: array
|
|
15980
15992
|
items: { $ref: '#/components/schemas/AllowRuleShadow' }
|
|
@@ -16192,6 +16204,24 @@ components:
|
|
|
16192
16204
|
`ASK_CLASSES` order, non-empty, `classes[0] === askClass`); one ask can belong to several, so read the whole
|
|
16193
16205
|
set, not just `askClass`. Present together with `askClass`; passed through verbatim in order (the server
|
|
16194
16206
|
neither sorts nor dedupes).
|
|
16207
|
+
boundGrantDigest:
|
|
16208
|
+
type: string
|
|
16209
|
+
description: >-
|
|
16210
|
+
server >= 7.118.0 (core 7.43.0 #1348 / #1357), "tool_approval" only — on a call of the engine's `Skill` tool: the hash of what
|
|
16211
|
+
approving it ALSO grants (the skill name + its two lists, canonical spelling); absent on every other card. Same value as the
|
|
16212
|
+
parked row's / the inbox row's twin (the live card and its parked twin reconcile by `(toolCallId, boundGrantDigest)`); passed
|
|
16213
|
+
through verbatim. A live "allow for this session" is remembered under "this tool + this digest" and answers only an ask with the
|
|
16214
|
+
same digest (same skill, same lists); a different skill or a changed list asks again. The durable `remember: "session"` never
|
|
16215
|
+
remembers a skill card.
|
|
16216
|
+
boundGrants:
|
|
16217
|
+
allOf: [{ $ref: '#/components/schemas/BoundSkillGrants' }]
|
|
16218
|
+
description: 'server >= 7.118.0, "tool_approval" only — the value `boundGrantDigest` hashes; present iff it is. DEPLOYMENT TEXT: render with contextual escaping.'
|
|
16219
|
+
classifierNoRuling:
|
|
16220
|
+
allOf: [{ $ref: '#/components/schemas/AutoModeFallbackCause' }]
|
|
16221
|
+
description: >-
|
|
16222
|
+
server >= 7.118.0 (core 7.43.0 #1353), "tool_approval" only, only on a `classifier_no_ruling` card — WHY the auto-mode classifier
|
|
16223
|
+
did not rule (`error` / `timeout` = it could not run; `parse_error` = it replied with no verdict), so a person decides. The server
|
|
16224
|
+
screens it with the engine's closed set (an out-of-set word is withheld). Display only.
|
|
16195
16225
|
denialLimitFallback: { $ref: '#/components/schemas/DenialLimitFallback' }
|
|
16196
16226
|
parked:
|
|
16197
16227
|
type: boolean
|
|
@@ -16388,13 +16418,11 @@ components:
|
|
|
16388
16418
|
promise and the per-leg index rules are stated verbatim on ToolApprovalFrame.ruleOffers.
|
|
16389
16419
|
persistedRuleShadowed:
|
|
16390
16420
|
type: string
|
|
16391
|
-
|
|
16392
|
-
|
|
16393
|
-
|
|
16394
|
-
|
|
16395
|
-
|
|
16396
|
-
(secret-redacted + length-bounded). UNTRUSTED for display. Semantics, the ABSENT != "you have no rule" clause and the deliberate separation
|
|
16397
|
-
from `governanceForced` are stated verbatim on ToolApprovalFrame.persistedRuleShadowed.
|
|
16421
|
+
description: >-
|
|
16422
|
+
server >= 7.13.0 (core 5.25.0 #144), ADDITIVE — the persisted rule this ask matched but which could not silence it; frame and
|
|
16423
|
+
card carry the SAME material (see ToolApprovalFrame.persistedRuleShadowed).
|
|
16424
|
+
RETIRED: server >= 7.117.0 (core 7.42.0) never mints this key on any face (its strict card schema even refuses a stored card
|
|
16425
|
+
carrying it); its successor is `allowRuleShadowed`. Kept for servers in [7.13.0, 7.117.0) only (sdk 16.3.0).
|
|
16398
16426
|
allowRuleShadowed:
|
|
16399
16427
|
type: array
|
|
16400
16428
|
items: { $ref: '#/components/schemas/AllowRuleShadow' }
|
|
@@ -16633,21 +16661,25 @@ components:
|
|
|
16633
16661
|
RuleOfferMatch:
|
|
16634
16662
|
type: string
|
|
16635
16663
|
description: >
|
|
16636
|
-
The MATCH FORM of a persisted rule — the server's `PERSISTED_RULE_MATCHES`, verbatim (the vocabulary
|
|
16664
|
+
The MATCH FORM of a persisted rule — the server's `PERSISTED_RULE_MATCHES`, verbatim and in order (the vocabulary
|
|
16637
16665
|
is OWNED BY CORE; the server only transcribes it and pins set equality, so core adding a member turns
|
|
16638
|
-
the server red at compile time).
|
|
16666
|
+
the server red at compile time). ONE vocabulary for two faces: a card's rule offer (`RuleOffer.match`, the
|
|
16667
|
+
server's card schema validates it with this very enum) and a persisted rule row (`GET /v1/rules`).
|
|
16639
16668
|
|
|
16640
16669
|
`exact` = this one command only (`Bash(git status)`). `prefix` = the historical colon-star
|
|
16641
16670
|
word-boundary prefix (`Bash(git status:*)`) and the ONLY spelling for a compound prefix form; it stays
|
|
16642
16671
|
a permanent compatibility read. `wildcard` = the trailing space-star form (`Bash(npm run *)`), the
|
|
16643
|
-
same predicate as `prefix` on a single command body
|
|
16644
|
-
|
|
16645
|
-
|
|
16646
|
-
|
|
16647
|
-
|
|
16648
|
-
|
|
16649
|
-
|
|
16650
|
-
|
|
16672
|
+
same predicate as `prefix` on a single command body. `subpath` = the `Read(//abs-dir/**)` directory form,
|
|
16673
|
+
which never holds a command. `path` (server >= 7.67.0) = the path-family pattern grammar for deny/ask/allow
|
|
16674
|
+
rules (`//abs`, `~/`, root-relative, cwd-relative; `*` within a segment, `**` across segments). `name` /
|
|
16675
|
+
`name_prefix` (server >= 7.118.0 / core 7.43.0 #1318) = the `Skill` name grammar: `Skill(<name>)` exact,
|
|
16676
|
+
`Skill(<prefix>:*)` a name prefix with NO word boundary; a skill card's "don't ask again" offer carries
|
|
16677
|
+
`match: "name"`. A replica rolled back to 7.117.x reads a `name` / `name_prefix` row as a damaged row.
|
|
16678
|
+
|
|
16679
|
+
🔴 SEVEN MEMBERS, closed: the server validates the member on both faces. A consumer type narrower than the
|
|
16680
|
+
mint point is a lie — it lets a `switch` claim exhaustiveness the wire does not honour (SDK <= 16 declared
|
|
16681
|
+
four here, so a `path` offer and every 7.118.0 skill-card offer were dropped as malformed).
|
|
16682
|
+
enum: [exact, prefix, wildcard, subpath, path, name, name_prefix]
|
|
16651
16683
|
|
|
16652
16684
|
RuleOfferUncoveredReason:
|
|
16653
16685
|
type: string
|
|
@@ -17223,25 +17255,66 @@ components:
|
|
|
17223
17255
|
additionalProperties: false
|
|
17224
17256
|
properties:
|
|
17225
17257
|
label: { type: string, maxLength: 256 }
|
|
17226
|
-
|
|
17227
|
-
|
|
17228
|
-
enum: [policy]
|
|
17258
|
+
decidedBy:
|
|
17259
|
+
allOf: [{ $ref: '#/components/schemas/ApprovalDecidedBy' }]
|
|
17229
17260
|
description: >-
|
|
17230
|
-
server
|
|
17231
|
-
|
|
17232
|
-
|
|
17233
|
-
|
|
17234
|
-
`
|
|
17235
|
-
|
|
17236
|
-
|
|
17237
|
-
|
|
17238
|
-
|
|
17239
|
-
|
|
17240
|
-
|
|
17241
|
-
|
|
17242
|
-
|
|
17243
|
-
|
|
17244
|
-
|
|
17261
|
+
server >= 7.118.0 — WHO made this decision (one key read the same way by all three decision doors; see
|
|
17262
|
+
ApprovalDecidedBy). ABSENT = `person`. This door's own answers (both refused before the decision, the card does
|
|
17263
|
+
not move): a `policy` allow on a person-required card ⇒ 400 `request.field_conflict`; outside the closed set ⇒
|
|
17264
|
+
400 `request.field_invalid`; any value on a direct-door worker ⇒ 400 `decided_by_not_in_proof`. An older server
|
|
17265
|
+
(no `approvalDecidedBy` bit) has a strict body here ⇒ 400 `request.body_shape`, nothing decided.
|
|
17266
|
+
|
|
17267
|
+
ApprovalDecidedBy:
|
|
17268
|
+
type: string
|
|
17269
|
+
enum: [person, policy]
|
|
17270
|
+
description: >
|
|
17271
|
+
server >= 7.118.0 (`APPROVAL_DECIDED_BY`, verbatim and in order) — WHO made an approval decision, one key on all three decision
|
|
17272
|
+
doors (live `POST /v1/tool-approvals/{approvalId}/respond`, durable `POST /v1/tasks/{taskId}/asks/{askId}/decision`, durable
|
|
17273
|
+
`POST /v1/approvals/{sessionId}/decide`). `person` = a person decided through the approval channel (ABSENT means this);
|
|
17274
|
+
`policy` = this deployment's own policy decided (a rule / posture answering on someone's behalf, a headless lane) — BOTH
|
|
17275
|
+
directions: the engine records `tool_end.gate.settlement.kind` = `policy_allowed` / `policy_refused` (`who.party: "policy"`).
|
|
17276
|
+
The only verb constraint is the engine's own: a person-required card (`requiresRealApproval`) cannot be ALLOWED by a policy.
|
|
17277
|
+
Send it only when `GET /v1/capabilities` carries `approvalDecidedBy: true`. Outside the closed set (including `sla_timeout` —
|
|
17278
|
+
"the wait ran out with nobody answering" is said only by the server's own deadline sweep — and the old `human` / `timeout`
|
|
17279
|
+
words) ⇒ 400 before the decision. It REPLACES the 7.93.3–7.117.x key `settledBy` (deny-only, the same fact), deleted
|
|
17280
|
+
by server 7.118.0 with no alias: sending `settledBy` ⇒ 400 `request.body_shape` on all three doors.
|
|
17281
|
+
`plan_review` does not take it (400 `request.body_shape`).
|
|
17282
|
+
|
|
17283
|
+
ApprovalReach:
|
|
17284
|
+
type: string
|
|
17285
|
+
enum: [none]
|
|
17286
|
+
description: >
|
|
17287
|
+
The closed set of the request key `approvalReach` (core 7.43.0 #1354 `APPROVAL_REACH`, verbatim): `none` — nobody on this run can be
|
|
17288
|
+
asked to approve. There is no positive word ("a person is reachable" is the `interactionPosture: "interactive"` promise).
|
|
17289
|
+
|
|
17290
|
+
AutoModeFallbackCause:
|
|
17291
|
+
type: string
|
|
17292
|
+
enum: [error, timeout, parse_error]
|
|
17293
|
+
description: >
|
|
17294
|
+
core 7.43.0 `AUTO_MODE_FALLBACK_CAUSES`, verbatim and in order — why the auto-mode classifier did not rule: `error` / `timeout` (it could
|
|
17295
|
+
not run), `parse_error` (it replied with no verdict). Carried as `classifierNoRuling` on a `classifier_no_ruling` card (server >= 7.118.0).
|
|
17296
|
+
|
|
17297
|
+
BoundSkillGrants:
|
|
17298
|
+
type: object
|
|
17299
|
+
description: >
|
|
17300
|
+
core 7.43.0 `BoundSkillGrants` (three keys, verbatim) — what approving an engine `Skill` card ALSO grants: `skill` = the loaded skill's
|
|
17301
|
+
name (`null` = the call names no loaded skill; it grants nothing), `allowedTools` = what activation lets be used without asking,
|
|
17302
|
+
`disallowedTools` = what activation refuses for the rest of the run (canonical spelling: de-duplicated, sorted). DEPLOYMENT TEXT.
|
|
17303
|
+
required: [skill, allowedTools, disallowedTools]
|
|
17304
|
+
additionalProperties: false
|
|
17305
|
+
properties:
|
|
17306
|
+
skill: { type: ['string', 'null'] }
|
|
17307
|
+
allowedTools: { type: array, items: { type: string } }
|
|
17308
|
+
disallowedTools: { type: array, items: { type: string } }
|
|
17309
|
+
|
|
17310
|
+
SteerReplayFacet:
|
|
17311
|
+
type: string
|
|
17312
|
+
enum: [text, trusted, priority, actor]
|
|
17313
|
+
description: >
|
|
17314
|
+
core 7.44.0 #1374 `STEER_REPLAY_FACETS`, verbatim and in order — the facets on which a replayed steering input id differs from the
|
|
17315
|
+
accepted one. Carried as `detail.differs` (a non-empty subset, the engine's order) on 409 `steering.duplicate_input_id` (server >= 7.118.0):
|
|
17316
|
+
contains `text` ⇒ a different instruction, reissue with a fresh key; no `text` ⇒ the SAME instruction — do NOT change the key (it
|
|
17317
|
+
would queue twice), replay with the original values.
|
|
17245
17318
|
|
|
17246
17319
|
AskDecisionAck:
|
|
17247
17320
|
type: object
|
|
@@ -18381,7 +18454,7 @@ components:
|
|
|
18381
18454
|
type: object
|
|
18382
18455
|
description: 'Approval seam: wired form, which seat it came from, and (effective half) what it adjudicates to.'
|
|
18383
18456
|
properties:
|
|
18384
|
-
form: { type: string, enum: [callback, allow, deny, absent] }
|
|
18457
|
+
form: { type: string, enum: [callback, allow, deny, absent, declared_none], description: '`declared_none` (server >= 7.118.0 / core 7.43.0): the run declared nobody on it can be asked (`approvalReach: "none"`), so a callback beside the declaration is not this run''s approver.' }
|
|
18385
18458
|
provenance: { type: string, enum: [spec, deps] }
|
|
18386
18459
|
effective: { type: string, enum: [human_reachable, auto_allow, auto_deny, park_only, unresolved] }
|
|
18387
18460
|
question:
|
|
@@ -19759,17 +19832,11 @@ components:
|
|
|
19759
19832
|
engine's catalogue, not on this wire. READ IT; do not switch over a closed list. A name the grammar
|
|
19760
19833
|
does not know is still a 400 on the input faces.
|
|
19761
19834
|
match:
|
|
19762
|
-
|
|
19763
|
-
|
|
19764
|
-
|
|
19765
|
-
|
|
19766
|
-
`
|
|
19767
|
-
compound prefix) · `wildcard` = the space-star form, the SAME predicate as `prefix` on a single
|
|
19768
|
-
command body · `subpath` = the directory (read-containment) form, which never admits a command ·
|
|
19769
|
-
`path` = NEW in server >= 7.67.0, the path-family pattern grammar for deny/ask/allow rules (`//abs`,
|
|
19770
|
-
`~/`, root-relative, cwd-relative; `*` within a segment, `**` across segments).
|
|
19771
|
-
SDK 8.7.0 and earlier declared only the first two — a consumer rendering this cell off a closed list
|
|
19772
|
-
must add the other three.
|
|
19835
|
+
allOf: [{ $ref: '#/components/schemas/RuleOfferMatch' }]
|
|
19836
|
+
description: >
|
|
19837
|
+
The match form — the same seven-member vocabulary as a card's rule offer (`RuleOfferMatch`, the server's
|
|
19838
|
+
`PERSISTED_RULE_MATCHES`): `exact` / `prefix` / `wildcard` / `subpath` / `path`, and from server >= 7.118.0
|
|
19839
|
+
`name` / `name_prefix` (a `Skill(<name>)` / `Skill(<prefix>:*)` rule).
|
|
19773
19840
|
command: { type: string, description: 'What the matcher actually compares against, parsed out of `rule` at mint time and stored so no consumer re-parses. Trust this + `match`, not your own parse of `rule`.' }
|
|
19774
19841
|
adds:
|
|
19775
19842
|
type: array
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@sema-agent/sdk",
|
|
3
|
-
"version": "
|
|
3
|
+
"version": "17.0.0",
|
|
4
4
|
"description": "Typed, zero-runtime-dependency SDK for the Sema agent fleet usage plane. The shared substrate for all doors (CC/Codex MCP façade + web). Server-side only — tokens never enter the browser.",
|
|
5
5
|
"type": "module",
|
|
6
6
|
"license": "BUSL-1.1",
|