[developer-review-engine-proof-v1] 1 — The packed runtime's public start and resume run this repository's executor, with halt, inspect, resume and cancel proved #1642

Activity

1. daniboomerang

   added this to the Productionize developer-review on Atta Engine milestone on Oct 25, 2026
2. daniboomerang

   added

   vinaya/tranche:developer-review-engine-proof-v1Tranche: developer-review-engine-proof-v1Tranche: developer-review-engine-proof-v1

   on Oct 25, 2026
3. daniboomerang commented on Oct 25, 2026

   

   daniboomerang

   on Oct 25, 2026

   MemberAuthor

   More actions

   Brief hash: d09349aaa0753fc825f3061a5b259b0c217c1df0002858db3e2dcb4f6b46f4f8
   For: [model] (coding-agent CLI on a dev machine, dispatched locally, unattended)
   Reason: Suggested agent-class — mid: one test against a documented public API plus three text corrections.
   Owner: the Principal
   Goal: The packed runtime's public start and resume run this repository's executor, with halt, inspect, resume and cancel proved
   Project: cli
   Tier: 1

   Closes #1642

   You are the AEG Developer. Run bun apps/cli/src/index.ts doctrine --role developer --print and read its output first. Mandatory.

   Objectives

   O1. A conformance test installs the packed Atta runtime, starts a steps-shaped Flow through the public startControlledRun with a caller-supplied node executor, halts it through public run control, reads its checkpoint, continues it through the public resumeControlledRun without re-running a finished step, and cancels a second run through public run control; every one of those calls is a public packed export.
   O2. The same Flow run under the built-in executor produces the same lifecycle event kinds in the same order and the same checkpoint transitions as under the caller-supplied executor; the test asserts the two sequences are equal.
   O3. An import check in that test refuses any import of the packed runtime that is not a public package export, so the run-control proof cannot pass through an internal path.
   O4. The Engine capability matrix and the run-control boundary registry entries no longer describe caller-supplied executor support on public start and resume as missing; they name the new test as the delivered evidence for that boundary.
   O5. The new test runs in continuous integration in exactly one shard.
   O6. The new test file contains the cutover-gate row id in its full registry form, cutover-gate:gate-16, in the name of the test that proves the public start and resume sequence, so the registry's evidence entry can bind gate 16 to this file.

   Documentation

   • apps/cli/specs/dev-review-evaluation-program.md — section 6.17, the run-control boundary's required tasks, grader and rule (O1, O2, O3)
   • apps/cli/specs/dev-review-engine-capability-matrix.md — the consumption-boundary paragraph that records the run-control gap (O4)
   • apps/cli/tests/conformance/packed-runtime-consumer.test.ts — the existing packed-consumer test this one extends in shape (O1)
   • harness/dev-process/engine-runtime/run-process.ts — how this repository already passes its executor to the public start and resume (O1)
   • harness/dev-process/engine-runtime/run-process.test.ts — the existing halt-then-resume test shape on one checkpointer (O1)

   Premises

   Checked against the code when this task Issue was cut, and again against the default branch at dispatch — facts, not claims.

   • harness/dev-process/engine-runtime/run-process.ts contains return await startControlledRun({
   • apps/cli/specs/dev-review-engine-capability-matrix.md contains Public start and resume do
   • apps/cli/tests/eval/registry/modules.json contains Gate-16 completion test: packed-runtime-public-run-control.test.ts
   • apps/cli/tests/eval/registry/shards/typed-sources.json contains is the gate-16 start/resume proof when Atta Engine delivers that public seam
   • apps/cli/tests/ci-shards/shard-3.txt contains tests/conformance/packed-runtime-consumer.test.ts

   2. Context — read before doing anything

   • Tranche: developer-review-engine-proof-v1, task 1, Issue [developer-review-engine-proof-v1] 1 — The packed runtime's public start and resume run this repository's executor, with halt, inspect, resume and cancel proved #1642. Branch task/developer-review-engine-proof-v1/1. Depends-on: —, Conflicts-with: —.
   • Read Issue [developer-review-engine-proof-v1] 1 — The packed runtime's public start and resume run this repository's executor, with halt, inspect, resume and cancel proved #1642 in full for the complete rationale — do not re-derive it.
   • Revision: rendered at b39fb00e8e327a1a9ba79de16fc84eeeb5154a25 — the checkout's HEAD equaled the remote default branch, and no pinned file below carried an uncommitted change, when these facts were read.
   • Boundary — Write the evidence that the packed runtime's public start and resume accept this repository's executor, and correct the documents that still say they do not. It does NOT change any runtime, Flow, action or executor code, and it does NOT re-pack the vendored tarballs. Pinned files: apps/cli/tests/conformance/packed-runtime-public-run-control.test.ts, apps/cli/tests/ci-shards/shard-3.txt, apps/cli/specs/dev-review-engine-capability-matrix.md, apps/cli/tests/eval/registry/modules.json, apps/cli/tests/eval/registry/shards/typed-sources.json. Out: harness/dev-process/engine-runtime/run-process.ts (read only); the CLI library directory is in the Surface only because a repository scanner there imports the pinned spec and registry files, and it is not edited; the vendored Engine tarballs and their digests; the gate-16 row's contract text, which already names its evidence by section and is not edited here.

   reader-resolvable-prose: code comments, READMEs and specs carry no Issue or pull-request numbers and no tranche slugs.

   doc-claims: every AEG:CLAIM line in a changed file is kept true in the same commit.

   • Traps to avoid — Do NOT import from dist/ paths or any subpath the package does not export; import only from the package names. Do NOT reuse a checkpointer across the two compared runs; give each run its own, as LangGraph's testing guidance requires. Do NOT spawn a real agent; the executor is a stub that records node ids, so the test runs the same with no Claude or Codex login. Do NOT edit the tarballs or their digests. A new test file must be listed in one apps/cli/tests/ci-shards/shard-N.txt in the same commit.

   3. Technical dependencies

   Dependency rationale — No Depends-on; no Conflicts-with.

   4. Technical surface map

   Create:

   • (none — every surface file already exists)

   Modify:

   • apps/cli/tests
   • apps/cli/specs
   • apps/cli/src/lib

   Out of surface: harness, processes, apps/log-server, aeg-root — a change there, or to any path no Surface in: glob covers, is still published: the review-loop driver records each such path in the pull request body's beyond-Surface block, and the reviewers judge it like any change.

   Premise pins

   Premise:

   • apps/cli/specs/dev-review-engine-capability-matrix.md sha256: 85665fc7b7ab32f721ce1bbfcea7c36f42f37fa16aaa8ade7ba8eba8a0caef02
   • apps/cli/tests/ci-shards/shard-3.txt sha256: e0a0a197a4d5d540b86fb40947e53a04edff0607413e09a3b52ec73a762c9b9d
   • apps/cli/tests/eval/registry/modules.json sha256: ded6c24d0b7470c2726772dce8fe89768e0b013cceea371e97ef30fc823c1fbc
   • apps/cli/tests/eval/registry/shards/typed-sources.json sha256: 15603ab943ea7b7affb02fcfd8973e0c2ba875e40a853ee4ed322aff14de5c74

   5. Pre-flight checks

   Step 0 (mandatory, verbatim):

   bun install --frozen-lockfile --silent
   

   1. Clean status; parent origin/main; branch suffix literal-matches the task id.
   2. Under the review loop (bun apps/cli/src/index.ts task run), the driver has already run the dispatch-readiness gate plus, on this repository toolchain, its own fuller derivation from its own unsandboxed process before dispatching you (re-derived at render time: it was ready), and dispatches you only when that verdict is READY — a NOT READY verdict refuses the turn before it starts, so being dispatched is the verdict: look for no file and run neither script, whose own gh call fails inside your sandbox. Working manually, with no driver behind you, fall back to roles/developer.md's own items 3/5/7 prose — each a complete, hand-run forge check in its own right.

   On any failure: STOP and report.

   6. Numbered parts — publish once with publish_changes when every Part is done, then open_pull_request

   1. Part 1 (O1, O3): The public start, halt, checkpoint read, resume and cancel sequence with a caller-supplied executor, and the public-export import check.

      Files (touches @attalabs/vinaya):

      • apps/cli/specs/dev-review-engine-capability-matrix.md
      • apps/cli/tests/ci-shards/shard-3.txt
      • apps/cli/tests/eval/registry/modules.json
      • apps/cli/tests/eval/registry/shards/typed-sources.json

      The pre-push hook runs the affected suite on your one push and refuses it on failure — do not run it yourself per Part.

   2. Part 2 (O2): The comparison of lifecycle events and checkpoint transitions against the built-in executor.

   3. Part 3 (O4): The capability matrix and both registry entries name the delivered evidence.

   4. Part 4 (O5): The shard entry.

   5. Part 5 (O6): The full gate-16 row id in the proving test's name.

   7. Documentation-update list

   .vinaya/doc-owners derivation matched zero bindings against this task's surface — no mechanically-derived doc updates. Confirm no additional doc artifact applies before treating this list as final.

   8. Verification before claiming done

   • Under the review loop, the controller runs the static gates: the commit hook applies safe format and lint fixes to staged files, typechecks affected packages, and runs the Vinaya checks; the pre-push hook runs the affected suite; run_checks and CI provide the authoritative results. Do not hand-run or paste static-gate commands. Working manually, run this repository's declared typecheck, lint/format, build, and documentation-gate commands yourself.
   • The pre-push hook runs the affected suite once, on your one push, and refuses it on failure; CI runs the full suite.
   • Every blast-radius consumer named in §4, re-verified by name.
   • The tier checklist in bun apps/cli/src/index.ts doctrine --role developer --print genuinely satisfied; use the controller's run_checks result, hooks, and CI as gate evidence.

   9. Test Plan

   cd apps/cli && bun test --timeout=30000 tests/conformance/packed-runtime-public-run-control.test.ts → 0 fail; start, halt, resume without replay and cancel pass with the caller's executor; built-in and caller sequences equal; the proving test's name contains `cutover-gate:gate-16`
   cd apps/cli && bun test --timeout=30000 tests/conformance/packed-runtime-consumer.test.ts tests/eval/exit-check.test.ts tests/eval/registry.test.ts → 0 fail
   cd apps/cli && bun test --timeout=30000 tests/ci-shards.test.ts → 0 fail
   bunx turbo run typecheck --affected → exits 0
   

   10. Stop conditions

   • The public start or resume refuses or ignores the caller-supplied executor in the installed package.
   • The built-in and caller-supplied runs differ in lifecycle events in a way the public types do not explain.

   Stop-and-escalate — Stop if the public startControlledRun or resumeControlledRun refuses or ignores the caller-supplied executor in the installed package; paste the failing assertion, because that means the boundary is not delivered and the documents must keep saying so. Stop if the built-in and caller-supplied runs differ in lifecycle events in a way the public types do not explain; paste both sequences.

   11. Constraints

   Autonomy: Do not stop to ask clarifying questions. For any ambiguity not covered by a Section 10 stop condition, choose the most reasonable option consistent with this brief, record the choice in the PR body at open, or in a PR comment after open, and continue. Halt only for the explicit Section 10 stop conditions — and when you halt, record the blocker in a PR comment or an Issue comment rather than waiting interactively for input.

   12. Deliverable

   • PR title (exact): [developer-review-engine-proof-v1] 1 — The packed runtime's public start and resume run this repository's executor, with halt, inspect, resume and cancel proved
   • Open the PR only via the open_pull_request tool (title and body); publish your commits with publish_changes and update the body with update_pull_request_body/refresh_evidence — you hold no gh/git push credential.
   • Publish once, with publish_changes, when every Part is done; then open the pull request with open_pull_request. Publish again only to answer review findings or a red gate — a refused hook, a failing check, a merge conflict.
   • PR body = the Developer's PR report (print it with bun apps/cli/src/index.ts doctrine --template pr-report --print), with Closes #1642 at the top of the header block. This brief stays frozen on the task Issue and never rides in the body.
   • Pre-open: tier checklist satisfied. The documentation gate is the controller's — run_checks and CI run it.
   • Include git diff origin/main...HEAD --stat as context only; Surface is judged by the driver's surface-scope check, never by this hand-run diff. Token use needs no report — the Vinaya log records it.
   • Then STOP. Review and Verification are separate invocations.
4. daniboomerang

   self-assigned this

   on Oct 25, 2026
5. daniboomerang commented on Oct 25, 2026

   

   daniboomerang

   on Oct 25, 2026

   MemberAuthor

   More actions

   The dev-review-loop paused: infrastructure — enforcement control 'local-gate' is inactive: ring 0 is INERT in this working copy — hooks are tracked at .vinaya/hooks but core.hooksPath is set to '~/Work/Repositories/Me/atta-labs/vinaya/.vinaya/hooks' (git config is never cloned). Remedy: Run git config core.hooksPath .vinaya/hooks once per cl….

   No pull request exists yet for this task, so the pause is recorded on this Issue instead.
   A Principal ruling is needed before this can continue. Post it as a comment on this Issue whose FIRST line is exactly this marker — the second number is the ordinal, and it must exceed the one this pause recorded:

   <!-- aeg:principal:ruling:1642-1 -->
   Your ruling text here.
   

   Then continue with:

   vinaya task run developer-review-engine-proof-v1 1 --agent claude --model opus --runner atta-engine
   

   The Operator's own task_resume clears this pause too, reading that same ruling from this Issue; task_cancel ends it instead.
6. daniboomerang commented on Oct 25, 2026

   

   daniboomerang

   on Oct 25, 2026

   MemberAuthor

   More actions

   The dev-review-loop paused: missing_source — required source(s) could not be prepared, so the Developer turn did not start: apps/cli/specs/dev-review-evaluation-program.md (missing_in_repository: apps/cli/specs/dev-review-evaluation-program.md does not exist at b39fb00); apps/cli/specs/dev-review-engine-capab….

   No pull request exists yet for this task, so the pause is recorded on this Issue instead.
   A Principal ruling is needed before this can continue. Post it as a comment on this Issue whose FIRST line is exactly this marker — the second number is the ordinal, and it must exceed the one this pause recorded:

   <!-- aeg:principal:ruling:1642-1 -->
   Your ruling text here.
   

   Then continue with:

   vinaya task run developer-review-engine-proof-v1 1 --agent claude --model claude-sonnet-5-5 --runner atta-engine
   

   The Operator's own task_resume clears this pause too, reading that same ruling from this Issue; task_cancel ends it instead.
7. daniboomerang commented on Oct 25, 2026

   

   daniboomerang

   on Oct 25, 2026

   MemberAuthor

   More actions

   The dev-review-loop paused: infrastructure — an uncaught error ended round 1's own processing: required source(s) could not be prepared, so the Developer turn did not start: apps/cli/specs/dev-review-evaluation-program.md (missing_in_repository: apps/cli/specs/dev-review-evaluation-program.md does not exist at b39fb00….

   No pull request exists yet for this task, so the pause is recorded on this Issue instead.
   A Principal ruling is needed before this can continue. Post it as a comment on this Issue whose FIRST line is exactly this marker — the second number is the ordinal, and it must exceed the one this pause recorded:

   <!-- aeg:principal:ruling:1642-1 -->
   Your ruling text here.
   

   Then continue with:

   vinaya task run developer-review-engine-proof-v1 1 --agent claude --model opus --runner atta-engine
   

   The Operator's own task_resume clears this pause too, reading that same ruling from this Issue; task_cancel ends it instead.
8. daniboomerang commented on Oct 25, 2026

   

   daniboomerang

   on Oct 25, 2026

   MemberAuthor

   More actions

   The dev-review-loop paused: infrastructure — an uncaught error ended round 1's own processing: the developer's prior launch (attempt 1) ended incomplete (no failure recorded) before its vendor session was ever bound — cannot resume the exact session, and starting a fresh one would lose worker continuity.

   No pull request exists yet for this task, so the pause is recorded on this Issue instead.
   A Principal ruling is needed before this can continue. Post it as a comment on this Issue whose FIRST line is exactly this marker — the second number is the ordinal, and it must exceed the one this pause recorded:

   <!-- aeg:principal:ruling:1642-1 -->
   Your ruling text here.
   

   Then continue with:

   vinaya task run developer-review-engine-proof-v1 1 --agent claude --model opus --runner atta-engine
   

   The Operator's own task_resume clears this pause too, reading that same ruling from this Issue; task_cancel ends it instead.
9. daniboomerang

   closed this as completedin #1662on Oct 25, 2026