@agentskit/harness 0.13.0 → 0.14.0

package/docs/ADR-0019-human-decision-attestation.md

@@ -13,7 +13,10 @@ Record `approval.recorded` and `authorization.recorded` events in the existing
 hash-chained lifecycle log. Each event contains the decision, resulting state,
 verification digest, source revision, contract hash, and human actor; tracking
 authorization also contains its declared target. Rejections are recorded too,
-so every human terminal decision is auditable.
+so every human terminal decision is auditable. The default tracking mode is
+goal-scoped: one human approval covers the declared tracking effect and emits
+both events. A project may set `tracking.authorization: "separate"` when an
+independent tracking decision is required.
 
 The existing freshness and verification-attestation checks run before either
 decision is accepted. The `run.json` fields remain a convenient projection,
@@ -22,6 +25,8 @@ but the event is the audit record.
 ## Consequences
 
 Consumers can independently prove which verified result was approved and which
-external tracking target was authorized. Legacy logs remain readable and are
-reported as legacy until they contain the new hashed event protocol. This is an
-audit attestation, not a digital signature or external notarization.
+external tracking target was authorized. Humans do not need to copy a run ID or
+digest: approval commands resolve the latest pending run, while the identifiers
+remain in the audit record. Legacy logs remain readable and are reported as
+legacy until they contain the new hashed event protocol. This is an audit
+attestation, not a digital signature or external notarization.

package/docs/MODULE-BOUNDARIES.md

@@ -164,7 +164,7 @@ How these seams are wired into the keep-pushing loop (and what is still only a k
 | --- | --- | --- | --- |
 | Doc Bridge | `src/adapters/doc-bridge.ts` | Reads a local index | Keep behind `ContextProvider`; measure context hit/quality separately. |
 | Orca | `src/adapters/orca.ts` | None; produces argv and lifecycle projections only | Keep lease/worktree/issue-lock/SHA planning provider-neutral; execution belongs to the orchestrator. |
-| Linear/GitHub/other tracker | `src/adapters/tracking.ts` callback | Caller-owned network mutation | Require idempotency key and explicit tracking authorization. |
+| Linear/GitHub/other tracker | `src/adapters/tracking.ts` callback | Caller-owned network mutation | Require idempotency key; goal approval covers declared tracking by default, with `tracking.authorization: "separate"` as an explicit opt-out. |
 | Orca CLI (loop) | `src/adapters/orca-cli.ts`, `src/adapters/linear-orca.ts` via `CommandRunner` | Read-only `--json` calls in the doctor; dispatch/mutation arrive in later loop phases | Argv only, never a shell string; every call bounded by a timeout; envelope `ok:false` fails closed. |
 | Coding-agent CLIs | `src/adapters/providers.ts` | PATH lookup and optional probe command | Env keys are names only; usage comes from Orca, never from provider SDKs. |
 | GitHub | `src/adapters/github-cli.ts` via `CommandRunner` | `gh pr view/list` (optionally `--label`), `gh api PUT …/merge` with `sha=<reviewed head>`, `gh pr comment`, `gh pr edit --remove-label` | Merge is refused by GitHub when the head moved; every call argv-based and bounded; label intake never calls the merge argv. |

package/loop.config.example.yaml

@@ -23,10 +23,33 @@ orca:
   minVersion: 1.4.200
   timeoutMs: 20000
 
+# Revisão mais rigorosa só nas fatias que pedem, por label. A revisão É o gate quando não há CI, e um
+# contrato que congela evidência não deve ser julgado com o mesmo orçamento de um ajuste de copy.
+# Primeira entrada que casa vence, e só os campos nomeados são trocados.
+# reviewOverrides:
+#   - anyLabels: [layer:L2, layer:L3]
+#     votes: 2
+#     minSeverity: nit
+#     reason: caminho crítico do MVP
+
+# Suítes já vermelhas na branch base. O harness NÃO roda o verifyCommand — o worker roda, na worktree
+# dele — então tolerar falha conhecida é informação no briefing, não parsing de saída. A issue de
+# rastreamento é obrigatória: quarentena sem dono vira permanente.
+# knownFailures:
+#   - path: packages/foo/tests/bar.test.ts
+#     issue: ENG-123
+#     reason: corrida de escrita em arquivo, pré-existente
+
 linear:
   workspaceId: <linear-workspace-id> # orca linear team list --workspace all --json
   teamKey: ENG
-  person: my-linear-display-name     # the ONLY queue this machine drains
+  person: my-linear-display-name     # the queue this machine drains (and, under unassigned ownership, the name it claims with)
+  queueOwnership: person             # person = the issues assigned to `person`. unassigned = the issues nobody holds,
+                                     # with the assignee as a transient claim the loop writes on dispatch and clears on
+                                     # return, so several machines share one priority-ordered queue.
+                                     # Switching to unassigned REQUIRES clearing the backlog's assignees: with `person`
+                                     # and an unassigned backlog the queue comes back empty and the loop looks healthy
+                                     # while doing nothing.
   people:                            # display name -> Linear user id (orca linear team members …)
     my-linear-display-name: <linear-user-id>
   rotation:
@@ -35,7 +58,11 @@ linear:
     advanceWhenEmpty: true           # never switch while an implementation lease remains
   states: [Todo, Ready]              # dispatchable states; anything already started is left alone
   excludeLabels: [blocked, needs-info]
-  requireLabels: []
+  requireLabels: []                  # ALL of these (AND); empty = no constraint
+  anyLabels: []                      # at least ONE of these (OR) — how a machine declares the slices it
+                                     # drains, e.g. [layer:L2, layer:L3]. requireLabels cannot say this:
+                                     # it demands every label on the same issue, so two layers there
+                                     # match nothing and the queue comes back silently empty.
   projects: []                       # empty = all projects
   order: [priority, updatedAt]       # urgent first; priority 0 (none) sorts last
   maxQueue: 50

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@agentskit/harness",
-  "version": "0.13.0",
+  "version": "0.14.0",
   "description": "Portable, evidence-backed development harness for coding agents.",
   "type": "module",
   "bin": {
@@ -87,20 +87,20 @@
     "prepublishOnly": "pnpm typecheck && pnpm build && pnpm test"
   },
   "dependencies": {
-    "commander": "^14.0.0",
+    "commander": "^14.0.3",
     "ink": "7.1.1",
     "react": "19.3.0",
-    "yaml": "^2.9.0",
-    "zod": "^4.6.2"
+    "yaml": "^2.9.1",
+    "zod": "^4.6.5"
   },
   "devDependencies": {
-    "@types/node": "^22.10.0",
+    "@types/node": "^22.20.3",
     "@types/react": "^19.3.0",
     "@vitest/coverage-v8": "4.1.11",
     "ink-testing-library": "4.0.0",
-    "tsup": "^8.5.0",
-    "typescript": "^5.7.2",
-    "vitest": "^4.1.10"
+    "tsup": "^8.5.1",
+    "typescript": "^5.9.3",
+    "vitest": "^4.1.11"
   },
   "license": "MIT",
   "packageManager": "pnpm@10.34.4",