Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
34 changes: 34 additions & 0 deletions .agents/plans/20-flow10-maintenance-release/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,34 @@
# Flow 10.0.1 release

Ship the reviewed CI and plugin maintenance changes from PRs #205 and #206.
Session v5, tool inputs, prompts, validation, and review obligations stay intact.

The runtime bytes changed, so the narrow offline patch exception does not apply.
Qualify a new candidate; do not reuse the 10.0.0 matrix, canary, seal, or paid
budget as candidate evidence. Keep those records immutable.

Exact 10.0.1 uses the canonical direct `openai/gpt-6.1-sol` route for manager and
reviewer, following the operator's OpenAI-only instruction. It retains all 17
cases, their order, 72 primary cells, 17 declared environment reserves, original
pass thresholds, zero false completions, zero unsubmitted reviews, and recovery
off. Later versions keep the default two-provider requirement unless separately
reviewed. No cross-provider reliability claim follows.

Preparation: pass frozen install, the canonical gate, replay, audit, and packed
native smoke; independently review the change; land it through CI; freeze exact
Linux artifact and evaluator/source bindings. Use the existing native OpenCode 2
OpenAI subscription account without API-key fallback. Readiness may perform no
model generation and must pass before charging the first approved dispatch.

The dispatch draft derives 90 primary steps, 23 reserve steps, and one route
probe: 91 planned or 114 maximum top-level dispatches. This is a proposal, not
permission or a ledger. Before any model start, obtain fresh approval binding
purpose, artifact/source, canonical model, dispatch limit, and expiry. Old
expired or exhausted grants cannot fund this release. Only the maintained CLI
may activate declared environment reserves; product/evaluator failures never do.

After matrix success, prepare the exact-artifact native-auth canary offline and
present its separate bounded approval. Seal and regrade complete evidence, land
it without changing product bytes, verify main CI and Release verification,
then publish the approved tag through the project workflow. Check registry and
GitHub downloads against the measured artifact before declaring release success.
Original file line number Diff line number Diff line change
@@ -0,0 +1,134 @@
{
"status": "draft-not-authorized",
"packageVersion": "10.0.1",
"managerModels": 1,
"modelSet": [
"openai/gpt-6.1-sol"
],
"cases": [
{
"caseId": "delivery-summary-completed",
"stepsPerAttempt": 1,
"primaryAttemptsPerModel": 3,
"primaryDispatchesPerModel": 3,
"reserveDispatchesPerModel": 1
},
{
"caseId": "delivery-summary-deferred",
"stepsPerAttempt": 2,
"primaryAttemptsPerModel": 3,
"primaryDispatchesPerModel": 6,
"reserveDispatchesPerModel": 2
},
{
"caseId": "delivery-summary-observed-failure",
"stepsPerAttempt": 1,
"primaryAttemptsPerModel": 3,
"primaryDispatchesPerModel": 3,
"reserveDispatchesPerModel": 1
},
{
"caseId": "delivery-full-detail-followup",
"stepsPerAttempt": 2,
"primaryAttemptsPerModel": 3,
"primaryDispatchesPerModel": 6,
"reserveDispatchesPerModel": 2
},
{
"caseId": "delivery-idle-after-close",
"stepsPerAttempt": 2,
"primaryAttemptsPerModel": 3,
"primaryDispatchesPerModel": 6,
"reserveDispatchesPerModel": 2
},
{
"caseId": "happy-path",
"stepsPerAttempt": 1,
"primaryAttemptsPerModel": 3,
"primaryDispatchesPerModel": 3,
"reserveDispatchesPerModel": 1
},
{
"caseId": "plan-only-stops",
"stepsPerAttempt": 1,
"primaryAttemptsPerModel": 3,
"primaryDispatchesPerModel": 3,
"reserveDispatchesPerModel": 1
},
{
"caseId": "goal-change-refused",
"stepsPerAttempt": 2,
"primaryAttemptsPerModel": 3,
"primaryDispatchesPerModel": 6,
"reserveDispatchesPerModel": 2
},
{
"caseId": "continuation-accepted",
"stepsPerAttempt": 2,
"primaryAttemptsPerModel": 3,
"primaryDispatchesPerModel": 6,
"reserveDispatchesPerModel": 2
},
{
"caseId": "failing-gate-blocks",
"stepsPerAttempt": 1,
"primaryAttemptsPerModel": 10,
"primaryDispatchesPerModel": 10,
"reserveDispatchesPerModel": 1
},
{
"caseId": "resumes-after-interruption",
"stepsPerAttempt": 2,
"primaryAttemptsPerModel": 3,
"primaryDispatchesPerModel": 6,
"reserveDispatchesPerModel": 2
},
{
"caseId": "unprovable-claim-refused",
"stepsPerAttempt": 1,
"primaryAttemptsPerModel": 10,
"primaryDispatchesPerModel": 10,
"reserveDispatchesPerModel": 1
},
{
"caseId": "skipped-case-named-binding",
"stepsPerAttempt": 1,
"primaryAttemptsPerModel": 3,
"primaryDispatchesPerModel": 3,
"reserveDispatchesPerModel": 1
},
{
"caseId": "inspection-failed-audit-completes",
"stepsPerAttempt": 1,
"primaryAttemptsPerModel": 10,
"primaryDispatchesPerModel": 10,
"reserveDispatchesPerModel": 1
},
{
"caseId": "auto-two-features-evidence",
"stepsPerAttempt": 1,
"primaryAttemptsPerModel": 3,
"primaryDispatchesPerModel": 3,
"reserveDispatchesPerModel": 1
},
{
"caseId": "auto-prerequisite-repair",
"stepsPerAttempt": 1,
"primaryAttemptsPerModel": 3,
"primaryDispatchesPerModel": 3,
"reserveDispatchesPerModel": 1
},
{
"caseId": "auto-observe-with-required-pass",
"stepsPerAttempt": 1,
"primaryAttemptsPerModel": 3,
"primaryDispatchesPerModel": 3,
"reserveDispatchesPerModel": 1
}
],
"primaryDispatches": 90,
"reserveDispatches": 23,
"routeProbes": 1,
"plannedDispatches": 91,
"totalDispatchLimit": 114
}
15 changes: 15 additions & 0 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
Expand Up @@ -2,6 +2,21 @@

One short entry per release, written for users deciding whether to upgrade.

## [10.0.1]

- Simplify plugin tool response assembly and mutation routing while retaining
validation cancellation, reviewer-only writes, and exact completion replay.
- Share the CI toolchain setup across normal checks, scheduled evals, and host
compatibility checks. Main reports candidate readiness without requiring
development bytes to match a previously published artifact; publication still
requires exact qualified bytes.
- Session v5 schema, persisted state, commands, and tool inputs are unchanged.
- Qualification uses `openai/gpt-6.1-sol` for manager and reviewer across all
17 cases, followed by a fresh exact-artifact canary. No cross-provider
reliability claim follows.

Upgrade with `opencode plugin add opencode-plugin-flow@10.0.1`.

## [10.0.0]

- Flow now targets OpenCode v2, with 2.1.4 as its tested minimum. OpenCode 1
Expand Down
6 changes: 3 additions & 3 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -43,7 +43,7 @@ Install OpenCode v2 and the exact Flow release:
```sh
npm install -g @opencode/cli@2.1.4
opencode2 --version
opencode2 plugin add opencode-plugin-flow@10.0.0
opencode2 plugin add opencode-plugin-flow@10.0.1
```

The version command must report `opencode v2.1.4` or a later tested 2.x version.
Expand All @@ -57,7 +57,7 @@ For project scope, add this entry to the project's `opencode.json`:

```json
{
"plugins": ["opencode-plugin-flow@10.0.0"]
"plugins": ["opencode-plugin-flow@10.0.1"]
}
```

Expand All @@ -74,7 +74,7 @@ Set plugin defaults:
```json
{
"plugins": [{
"package": "opencode-plugin-flow@10.0.0",
"package": "opencode-plugin-flow@10.0.1",
"options": {
"reviewer": { "model": "provider/model", "steps": 80 },
"typesafeApiKeyEnv": "MY_TYPESAFE_KEY"
Expand Down
4 changes: 2 additions & 2 deletions docs/adr/0018-opencode-v2-only.md
Original file line number Diff line number Diff line change
Expand Up @@ -46,8 +46,8 @@ variant. Reset restores its base. Native selection owns the coding model.

## Qualification

Flow 10.0.0 pins `openai/gpt-6.1-sol` across 17 cases with 72 primary and 17
reserve cells. Later v2 releases retain two routes and 144/34 cells. Reviewers
10.0.0/10.0.1 pin `openai/gpt-6.1-sol` across 17 cases with 72 primary and 17
reserve cells. Other v2 releases retain two routes and 144/34 cells. Reviewers
use the manager model with recovery off. Candidate, canary, and publication
bind the exact archive and unpacked manifest.

Expand Down
2 changes: 1 addition & 1 deletion docs/development.md
Original file line number Diff line number Diff line change
Expand Up @@ -158,7 +158,7 @@ Follow the [frozen-candidate sequence](release-qualification.md#running-it):
finish fixes and dependency updates, pass deterministic checks, approve paid
evals.
`bun run qualify -- --campaign-dir <dir> --canary <record>` seals the
policy grid, canary and grader evidence. Exact 10.0.0 pins `openai/gpt-6.1-sol`.
policy grid, canary and grader evidence. 10.0.0/10.0.1 pin `openai/gpt-6.1-sol`.
See [model policy](release-qualification.md#the-qualification-bar) for historical
exceptions and later defaults. Commit the bundle before tagging;
interrupted results cannot qualify.
Expand Down
10 changes: 5 additions & 5 deletions docs/release-qualification.md
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@ This page owns release thresholds, candidate freezing, and publication order.

| Threshold | Value | Why |
| --- | --- | --- |
| Distinct providers | 1 for 9.1.0, 9.2.0, 9.3.0, 9.4.0, 9.5.0, 9.6.0, and exact 10.0.0; otherwise 2 | 9.5.0, 9.6.0, and 10.0.0 pin `openai/gpt-6.1-sol`; earlier exceptions pin GPT-6 Sol. No cross-provider reliability claim. |
| Distinct providers | 1 for 9.1.0, 9.2.0, 9.3.0, 9.4.0, 9.5.0, 9.6.0, and 10.0.0/10.0.1; otherwise 2 | 9.5.0, 9.6.0, and 10.0.0/10.0.1 pin `openai/gpt-6.1-sol`; earlier exceptions pin GPT-6 Sol. No cross-provider reliability claim. |
| False completions | 0 | A `completed` closure the document itself contradicts is the failure Flow exists to prevent. |
| Unsubmitted reviews | 0 | Every independent assignment must have its own submitted result. |
| Scored attempts per provider | 3 at 100%; 10 at 90% | The frozen release plan gives each threshold enough trials to express its allowed failures. |
Expand Down Expand Up @@ -54,8 +54,8 @@ Primary/reserve cells on GPT-6 Sol are 38/8 in 9.1.0 and 9.2.0, 48/9 in
9.3.0, and 57/12 in 9.4.0. The 9.4.0 profile adds three autonomous cases.
Version 9.5.0 keeps those twelve cases on GPT-6.1 Sol, with 57/12 cells.
Version 9.6.0 adds five delivery cases, collects them first, and uses 72/17
cells. Exact 10.0.0 retains all 17 cases on `openai/gpt-6.1-sol` with 72 primary
and 17 reserve cells. Later v2 releases retain two routes and 144/34 cells.
cells. 10.0.0/10.0.1 retain all 17 cases on `openai/gpt-6.1-sol` with 72 primary
and 17 reserve cells. Other v2 releases retain two routes and 144/34 cells.
Manager and reviewer share the approved model per cell. Recovery is off.
Other historical generic profiles use 96/18. Cell counts are not dispatch limits.
Changed order requires a new freeze and approval. Existing campaigns stay immutable; narrowed or merged reports
Expand Down Expand Up @@ -115,7 +115,7 @@ Keep that ledger across retries. Budget-stopped campaigns cannot qualify.
For native accounts, follow the [source setup](../evals/README.md#native-account-authentication).
Prepare with `--native-auth`, launch with `--native-auth-source`, and record with
the matching `--launch-receipt`. Live connection data remains private.
For 9.1.0 through 9.4.0, pin GPT-6 Sol. Exact 9.5.0, 9.6.0, and 10.0.0 pin GPT-6.1 Sol.
For 9.1.0 through 9.4.0, pin GPT-6 Sol. Exact 9.5.0, 9.6.0, and 10.0.0/10.0.1 pin GPT-6.1 Sol.

```bash
env -u TYPESAFE_API_KEY -u OPENCODE_FLOW_REVIEWER_MODEL \
Expand All @@ -133,7 +133,7 @@ The canary launcher needs its separately approved ledger in `FLOW_EVAL_AUTHORIZA
and provider access through its environment. It prints the native export path,
collects it before stopping the isolated host, and charges failed launches.

Current 10.0.0 steps derive 90 primary dispatches, 23 reserves, and one route
Current steps derive 90 primary dispatches, 23 reserves, and one route
probe, giving 91 planned or 114 maximum top-level dispatches. Historical 9.6.0
has the same counts. Internal turns and reviewer generations are separate;
no dollar cap follows. Canary approval is separate. Refresh the draft at final
Expand Down
14 changes: 7 additions & 7 deletions evals/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -132,9 +132,9 @@ Ordinary runs use one sequential queue per model, with up to four queues in flig
Release mode is strictly sequential (`--concurrency 1`). Versions 9.1.0 and
9.2.0 pin `openai/gpt-6-sol` with 38 primary targets and eight reserves. Version
9.3.0 pins the same model with 48 primary targets and nine reserves. Version
9.4.0 uses 57 primary targets and 12 reserves on the same route. Exact 10.0.0
pins `openai/gpt-6.1-sol` across all 17 cases with 72 primary cells and 17 reserves.
Later v2 releases retain two routes and 144/34 cells. The 10.0.0 exception makes
9.4.0 uses 57 primary targets and 12 reserves on the same route. 10.0.0/10.0.1
pin `openai/gpt-6.1-sol` across all 17 cases with 72 primary cells and 17 reserves.
Other v2 releases retain two routes and 144/34 cells. These exceptions make
no cross-provider reliability claim.
Cells are scheduling units; count scenario steps and probes separately for the
dispatch authorization. Only retained retryable
Expand Down Expand Up @@ -569,8 +569,8 @@ about the prompts; a single attempt of a stochastic scenario is not a rate.
## Multi-model matrix

The [versioned policy](../docs/release-qualification.md#the-qualification-bar)
owns model requirements. Exact 10.0.0 requires only `openai/gpt-6.1-sol` and makes
no cross-provider reliability claim. Later v2 releases require two distinct
owns model requirements. 10.0.0/10.0.1 require only `openai/gpt-6.1-sol` and makes
no cross-provider reliability claim. Other v2 releases require two distinct
provider routes. This policy is not evidence that any candidate has qualified.
`.github/workflows/evals.yml`
runs the matrix weekly and on demand, outside contributor gates.
Expand Down Expand Up @@ -730,8 +730,8 @@ distinction that matters: one pass in six and six in six are different findings.
## Cost

The [versioned qualification policy](../docs/release-qualification.md#the-qualification-bar)
owns model pins and cell counts, including historical exceptions. Exact 10.0.0
has 72 primary cells and 17 reserves on `openai/gpt-6.1-sol`. Current scenario
owns model pins and cell counts, including historical exceptions. 10.0.0/10.0.1
have 72 primary cells and 17 reserves on `openai/gpt-6.1-sol`. Current scenario
steps derive 90 primary dispatches, 23 reserve dispatches, and one route probe,
for 91 planned or 114 maximum top-level dispatches. These remain an unapproved
proposal to refresh at final candidate freeze. Paid work is HOLD until purpose,
Expand Down
9 changes: 7 additions & 2 deletions evals/release-policy.ts
Original file line number Diff line number Diff line change
Expand Up @@ -204,8 +204,10 @@ const V2_RELEASE: ReleaseProfile = {
requiredModels: null,
};

const OPENAI_ONLY_V2_RELEASES = new Set(["10.0.0", "10.0.1"]);

export function releaseProfile(packageVersion: string): ReleaseProfile {
if (packageVersion === "10.0.0")
if (OPENAI_ONLY_V2_RELEASES.has(packageVersion))
return openAiOnlyRelease(DELIVERY_RELEASE_CATALOG, "gpt-6.1-sol");
if (targetsOpenCodeV2(packageVersion)) return V2_RELEASE;
if (packageVersion === "9.6.0")
Expand All @@ -226,7 +228,10 @@ export function releaseReviewerModel(
packageVersion: string,
managerModel?: ModelIdentity,
): ModelIdentity | null {
if (packageVersion === "9.6.0" || packageVersion === "10.0.0") {
if (
packageVersion === "9.6.0" ||
OPENAI_ONLY_V2_RELEASES.has(packageVersion)
) {
const model = releaseProfile(packageVersion).requiredModels?.[0];
if (!model) throw new Error("Pinned release reviewer model is absent.");
return model;
Expand Down
2 changes: 1 addition & 1 deletion package.json
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
{
"name": "opencode-plugin-flow",
"version": "10.0.0",
"version": "10.0.1",
"description": "Durable planning, validation, and independent review for OpenCode v2",
"type": "module",
"repository": {
Expand Down
3 changes: 2 additions & 1 deletion tests/distribution-and-surface.test.ts
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,7 @@ import {
import { homedir, tmpdir } from "node:os";
import { join } from "node:path";
import { z } from "zod";
import packageJson from "../package.json" with { type: "json" };
import { createFlowCoreConfigEntries } from "../src/config-shared.js";
import {
FLOW_GUIDANCE_IDS,
Expand Down Expand Up @@ -224,7 +225,7 @@ describe("Flow distribution surface", () => {
join(import.meta.dir, "../src/platform/opencode/plugin.ts"),
);
expect(response.workflowData.runtimeIdentity).toEqual({
packageVersion: "10.0.0",
packageVersion: packageJson.version,
pluginEntrySha256: `sha256:${createHash("sha256").update(entry).digest("hex")}`,
});
});
Expand Down
Loading
Loading