@kontextmind/kxm 0.7.185 → 0.7.187

package/.claude-plugin/marketplace.json

@@ -11,7 +11,7 @@
       "name": "kxm",
       "source": "./plugins/kxm",
       "description": "Durable workflows, peer agents, and kxm tui",
-      "version": "0.7.185",
+      "version": "0.7.187",
       "category": "development",
       "tags": ["kxm", "multi-agent", "workflows", "mcp"]
     }

package/.kxm/README.md

@@ -37,7 +37,7 @@ Ids in this checkout follow one convention per kind:
 |---|---|---|
 | Role | The purpose: `planner`, `writer`, `reviewer-arch`, `reviewer-cli` | Unchanged |
 | Agent | The same id as its `role` | `planner`, `writer`, `reviewer-arch`, `reviewer-cli` |
-| Route | `<harness>-<model-slug>[-<provider>]`, with `.` written as `-` | `grok-grok-4-7`, `claude-fable`, `pi-qwen3-coder-plus-openrouter` |
+| Route | `<harness>-<model-slug>[-<provider>]`, with `.` written as `-` | `grok-grok-4-7`, `claude-fable`, `pi-qwen3-7-plus-openrouter` |
 | Workflow | `default`, `land`, or `<role>-only` | `default`, `land`, `writer-only`, `reviewer-arch-only`, `reviewer-cli-only` |
 | Agent step | The agent's role id | `writer`, `reviewer-arch`, `reviewer-cli` |
 

package/.kxm/roles/writer.yaml

@@ -10,7 +10,7 @@ description: Primary implementation agent.
 roster:
   - route: grok-grok-4-7
     effort: medium
-  - route: pi-qwen3-coder-plus-openrouter
+  - route: pi-qwen3-7-plus-openrouter
     effort: medium
   - route: agy-gemini-4-argon-high
     effort: medium

package/CHANGELOG.md

@@ -6,6 +6,15 @@ All notable user-facing changes are documented here. The project follows [Semant
 
 ### Added
 
+- **OpenRouter model prices and the inventory refresh weekly.**
+  `.github/workflows/refresh-models.yml` runs Monday at 13:00 UTC, or on
+  demand, and opens one pull request on `automation/model-refresh` when
+  `.kxm/prices.yaml` or the OpenRouter slice of `.kxm/models/inventory.yaml`
+  changes. A model still named by `.kxm/routes.yaml` or a roster is kept when
+  it leaves the catalog, and the `model-catalog-guard` check fails until a
+  human decides. Run it locally with
+  `node --disable-warning=ExperimentalWarning --experimental-strip-types scripts/refresh-model-catalog.mjs`.
+
 - **The witnessed MG1 Python control-plane foundation is now reproducible from this repository.**
   `python/kxm-control` provides the pinned Python 3.13 uv workspace, fixed
   `kxmd` Postgres/Temporal caller, Authentik user and runner verification
@@ -216,6 +225,15 @@ All notable user-facing changes are documented here. The project follows [Semant
 
 ### Changed
 
+- **Writer #2 is OpenRouter Qwen3.7 Plus.** `pi-qwen3-7-plus-openrouter`
+  (`openrouter/qwen/qwen3.7-plus`) replaces `pi-qwen3-coder-plus-openrouter`.
+  OpenRouter no longer lists `qwen/qwen3-coder-plus`. The live catalog on
+  2026-10-17 prices Qwen3.7 Plus at $0.32 input and $1.28 output per 1M
+  tokens, with a 1,000,000-token context and tool support. The old model
+  file stays `status: retired`, off every roster, and `qwen-openrouter-pi`
+  no longer resolves. `.kxm/prices.yaml` keeps the last Coder Plus row and
+  adds the Qwen3.7 Plus base tier. The dev helper's only Pi writer is
+  `openrouter/qwen/qwen3.7-plus`.
 - **The CLI critic is Codex GPT-6.1 Sol.** `codex-gpt-6-1-sol` is the
   `reviewer-cli` primary (`codex` / `gpt-6.1-sol`, effort `low`).
   `pi-qwen3-8-flash-openrouter` stays the other-vendor fallback, and

package/docs/contributing/assignment-runner.md

@@ -34,7 +34,7 @@ admits each route for specific roles and permissions:
 
 | Role | Admitted route (harness / model) | Vendor | Permission |
 |---|---|---|---|
-| `writer` | `grok` / `grok-4.7`; relief: `pi` / `openrouter/qwen/qwen3-coder-plus`; `agy` / `gemini-4-argon-high`, then `gemini-4-argon-medium`, then `gemini-3.8-flash-medium`, then `gemini-3.8-flash-high` | `xai`; `alibaba`; `google` | `edit` |
+| `writer` | `grok` / `grok-4.7`; relief: `pi` / `openrouter/qwen/qwen3.7-plus`; `agy` / `gemini-4-argon-high`, then `gemini-4-argon-medium`, then `gemini-3.8-flash-medium`, then `gemini-3.8-flash-high` | `xai`; `alibaba`; `google` | `edit` |
 | `planner` | `claude` / `fable` | `anthropic` | `read-only` |
 | `reviewer-arch` | `claude` / `fable` | `anthropic` | `read-only` |
 | `reviewer-cli` | `codex` / `gpt-6.1-sol`; fallback `pi` / `openrouter/qwen/qwen3.8-flash`, then `codex` / `gpt-5.6-sol` | `openai`; `alibaba`; `openai` | `read-only` |
@@ -347,7 +347,7 @@ and exits with status 2. It refuses when:
   segment must not be a braked native vendor (`x-ai`, `moonshotai` and
   `google-ai` count as `xai`, `moonshot` and `google`);
 - a Pi `edit` request comes from a role other than `experiment` or `writer`, or
-  a Pi writer asks for anything but `openrouter/qwen/qwen3-coder-plus` with
+  a Pi writer asks for anything but `openrouter/qwen/qwen3.7-plus` with
   `edit`;
 - the request asks for hooks or skills, sets `max_cost_usd` for a harness other
   than Claude, or sets `max_turns` for a harness other than Grok;

package/docs/contributing/ci-and-release.md

@@ -46,6 +46,7 @@ together.
 | Push to `main`, or manual | `ci.yml` (Validate matrix) | `validate:pr` on Linux and Windows for Node 22.19.0 and Node 24; not on a pull request |
 | Code pull request and code push | `ci.yml` (Plugin validation) | `claude plugin validate --strict` on the marketplace and the plugin |
 | Daily at 04:00 UTC, or manual | `nightly.yml` | `test:coverage:complete`, `check`, `check:generated`, `npm pack --dry-run` |
+| Monday at 13:00 UTC, or manual | `refresh-models.yml` | Fetches the public OpenRouter catalog, and opens or updates one pull request on `automation/model-refresh` when `.kxm/prices.yaml` or `.kxm/models/inventory.yaml` changes. Uses `GITHUB_TOKEN` only, then dispatches `ci.yml` on that branch |
 | Merged pull request | `auto-release.yml` | Tags the merge commit and dispatches `release.yml` |
 | Tag push or dispatch | `release.yml` | Verifies, packs and publishes (see [Release flow](#release-flow)) |
 | Manual only | `smoke.yml` | Real Pi smoke, currently disabled (see [Smoke tests](#smoke-tests)) |
@@ -269,7 +270,7 @@ The real Pi smoke calls paid models, so it never runs by default. Its workflow,
 `KXM_SMOKE_RUNNER` names a runner with Pi credentials. Both smoke models run as
 long-lived Pi workers, so each must pass the Pi native-vendor brake: pick two
 non-native routes such as
-`openrouter/qwen/qwen3-coder-plus,openrouter/z-ai/glm-5.3-flash`, never `xai/…`
+`openrouter/qwen/qwen3.7-plus,openrouter/z-ai/glm-5.3-flash`, never `xai/…`
 or `anthropic/…`. The container smoke needs
 Docker, a `pass-cli` session and a local Pi model store. It keeps secrets in a
 mode-600 file inside a temporary directory; set `KXM_SMOKE_VAULT` to read keys

package/docs/contributing/development.md

@@ -400,6 +400,28 @@ authentication are the protocol. A change to any of them needs:
 Prefer additive changes. KXM does not keep backward-compatible aliases for
 retired names: an old name fails closed instead.
 
+## Refresh model prices and inventory
+
+OpenRouter list prices and the model inventory are refreshed by
+`.github/workflows/refresh-models.yml` every Monday at 13:00 UTC. To run the
+same refresh before that, from the repository root:
+
+```bash
+node --disable-warning=ExperimentalWarning --experimental-strip-types scripts/refresh-model-catalog.mjs --dry-run
+```
+
+`--dry-run` prints models added, removed, and repriced, and writes nothing.
+Drop `--dry-run` to update `.kxm/prices.yaml` and `.kxm/models/inventory.yaml`.
+Exit 2 means a model named by `.kxm/routes.yaml` or a roster left the OpenRouter
+catalog; the refresh keeps that inventory row and sets `missingUpstream: true`.
+Review the diff and commit it yourself, or wait for the Monday pull request on
+`automation/model-refresh`. The field reference is
+[Weekly OpenRouter refresh](../reference/config-reference.md#weekly-openrouter-refresh).
+
+`kxm models inventory-refresh` still rewrites the whole inventory from Pi, Grok,
+agy, OpenRouter, and Nous, including a new `fetchedAt` timestamp. The weekly
+script does not call those local CLIs.
+
 ## Troubleshooting
 
 | Symptom | Cause | Fix |

package/docs/contributing/harness-routing-internals.md

@@ -9,7 +9,7 @@ This page records how the KXM repository applies [harness routing](../reference/
 
 The files use one id shape per kind. A route id is `<harness>-<model-slug>[-<provider>]`, with `.` written as `-`. An agent id equals its role. A workflow id is `default`, `land`, or `<role>-only`. An agent step id equals that role.
 
-This checkout's routes are `grok-grok-4-7`, `pi-qwen3-coder-plus-openrouter`, `agy-gemini-4-argon-high`, `agy-gemini-4-argon-medium`, `agy-gemini-3-8-flash-medium`, `agy-gemini-3-8-flash-high`, `claude-fable`, `codex-gpt-6-1-sol`, `codex-gpt-5-6-sol`, `pi-qwen3-8-flash-openrouter`, and `pi-glm-5-3-flash-openrouter`. Agents are `planner`, `writer`, `reviewer-arch`, and `reviewer-cli`. `opus-claude` was removed because `opus` is not an admitted selector. `qwen-token-plan/*` and `zai-coding-cn/*` left the admitted list because those harnesses are not allowlisted; their `prices.yaml` rows stay. Old ids resolve as aliases. Every roster entry names an effort. `reviewer-arch` dispatches `claude-fable`. #343's architecture critic ran on opus because the justfile recipe review-arch hardcoded that model; the role file on that commit listed only `fable-claude`. The helper now refuses `opus`.
+This checkout's admitted routes are `grok-grok-4-7`, `pi-qwen3-7-plus-openrouter`, `agy-gemini-4-argon-high`, `agy-gemini-4-argon-medium`, `agy-gemini-3-8-flash-medium`, `agy-gemini-3-8-flash-high`, `claude-fable`, `codex-gpt-6-1-sol`, `codex-gpt-5-6-sol`, `pi-qwen3-8-flash-openrouter`, and `pi-glm-5-3-flash-openrouter`. `pi-qwen3-coder-plus-openrouter` stays on disk with `status: retired` because OpenRouter removed `qwen/qwen3-coder-plus` (catalog 404); it is not admitted and not on a roster, and `qwen-openrouter-pi` does not resolve. Agents are `planner`, `writer`, `reviewer-arch`, and `reviewer-cli`. `opus-claude` was removed because `opus` is not an admitted selector. `qwen-token-plan/*` and `zai-coding-cn/*` left the admitted list because those harnesses are not allowlisted; their `prices.yaml` rows stay. Other old ids resolve as aliases. Every roster entry names an effort. `reviewer-arch` dispatches `claude-fable`. #343's architecture critic ran on opus because the justfile recipe review-arch hardcoded that model; the role file on that commit listed only `fable-claude`. The helper now refuses `opus`.
 
 ## This checkout's routes and roster
 
@@ -25,7 +25,7 @@ skills: []
 roster:
   - route: grok-grok-4-7
     effort: medium
-  - route: pi-qwen3-coder-plus-openrouter
+  - route: pi-qwen3-7-plus-openrouter
     effort: medium
   - route: agy-gemini-4-argon-high
     effort: medium
@@ -37,7 +37,7 @@ roster:
     effort: medium
 ```
 
-The writer role lists route ids, in order: `grok-grok-4-7`, then `pi-qwen3-coder-plus-openrouter`, then `agy-gemini-4-argon-high`, `agy-gemini-4-argon-medium`, `agy-gemini-3-8-flash-medium`, and `agy-gemini-3-8-flash-high`. `grok-grok-4-7` is harness `grok`, model `grok-4.7`, vendor `xai`. An agent file does not declare `harness:`. Argon is writer #3 and #4. Flash medium is the safety net because Argon needs a paid Google AI Pro, Ultra, or Enterprise plan and `agy -p` exits when the pinned model is missing. Flash high stays on the roster so that admitted model file is still used. The planner primary stays `claude-fable`. The CLI critic primary is `codex-gpt-6-1-sol`, then `pi-qwen3-8-flash-openrouter`, then `codex-gpt-5-6-sol`. The architecture critic primary stays `claude-fable`, then `pi-glm-5-3-flash-openrouter`. Argon can serve as a planner or architecture-review fallback; it is not on those rosters. GPT-6.1 Sol Pro is the same model with reasoning mode `pro`. The route id does not carry a reasoning mode, role effort has no `pro` value, and codex is limited to `reviewer-cli`, so that mode is not an architecture fallback.
+The writer role lists route ids, in order: `grok-grok-4-7`, then `pi-qwen3-7-plus-openrouter`, then `agy-gemini-4-argon-high`, `agy-gemini-4-argon-medium`, `agy-gemini-3-8-flash-medium`, and `agy-gemini-3-8-flash-high`. `grok-grok-4-7` is harness `grok`, model `grok-4.7`, vendor `xai`. An agent file does not declare `harness:`. Argon is writer #3 and #4. Flash medium is the safety net because Argon needs a paid Google AI Pro, Ultra, or Enterprise plan and `agy -p` exits when the pinned model is missing. Flash high stays on the roster so that admitted model file is still used. The planner primary stays `claude-fable`. The CLI critic primary is `codex-gpt-6-1-sol`, then `pi-qwen3-8-flash-openrouter`, then `codex-gpt-5-6-sol`. The architecture critic primary stays `claude-fable`, then `pi-glm-5-3-flash-openrouter`. Argon can serve as a planner or architecture-review fallback; it is not on those rosters. GPT-6.1 Sol Pro is the same model with reasoning mode `pro`. The route id does not carry a reasoning mode, role effort has no `pro` value, and codex is limited to `reviewer-cli`, so that mode is not an architecture fallback.
 
 `node scripts/kxm.mjs routes list`:
 
@@ -49,7 +49,7 @@ admitted google/gemini-4-argon-high
 admitted google/gemini-4-argon-medium
 admitted openai/gpt-5.6-sol
 admitted openai/gpt-6.1-sol
-admitted openrouter/qwen/qwen3-coder-plus
+admitted openrouter/qwen/qwen3.7-plus
 admitted openrouter/qwen/qwen3.8-flash
 admitted openrouter/z-ai/glm-5.3-flash
 admitted xai/grok-4.7
@@ -61,14 +61,14 @@ reviewer-cli codex-gpt-6-1-sol
 reviewer-cli pi-qwen3-8-flash-openrouter
 reviewer-cli codex-gpt-5-6-sol
 writer grok-grok-4-7
-writer pi-qwen3-coder-plus-openrouter
+writer pi-qwen3-7-plus-openrouter
 writer agy-gemini-4-argon-high
 writer agy-gemini-4-argon-medium
 writer agy-gemini-3-8-flash-medium
 writer agy-gemini-3-8-flash-high
 ```
 
-`routes admit --model openrouter/qwen/qwen3-coder-plus --dry-run` exits 2 with `select a model from the refreshed inventory` even though that selector is already admitted, because the inventory never carries `openrouter/…` selectors. Selectors like these are admitted by a Git-reviewed edit of `.kxm/routes.yaml`.
+`routes admit --model openrouter/qwen/qwen3.7-plus --dry-run` exits 2 with `select a model from the refreshed inventory` even though that selector is already admitted, because the inventory never carries `openrouter/…` selectors. Selectors like these are admitted by a Git-reviewed edit of `.kxm/routes.yaml`.
 
 On the capture machine, `kxm harness list` showed `claude` detected but logged out (`dispatch no (not_authenticated)`), `deepseek` not installed, and `grok`, `codex`, `kimi` and `agy` ready.
 
@@ -85,9 +85,9 @@ The issue-127 runner (`kxm assign`, see the [assignment runner](assignment-runne
     harness: grok
     model: grok-4.7
     vendor: xai
-  pi-qwen3-coder-plus-openrouter:
+  pi-qwen3-7-plus-openrouter:
     harness: pi
-    model: openrouter/qwen/qwen3-coder-plus
+    model: openrouter/qwen/qwen3.7-plus
     vendor: alibaba
 ```
 
@@ -108,7 +108,7 @@ The product layers are in [What the brake refuses](../reference/harness-routing.
 
 | Layer | Where | What it refuses | What you see |
 |---|---|---|---|
-| Dev helper | `scripts/harness-run.mjs` | A braked Pi provider, a Pi provider outside the allowlist, a Pi writer other than `openrouter/qwen/qwen3-coder-plus`, or an aggregator id whose vendor segment is a native vendor | `pi brake: xai has a native harness; refusing Pi impersonation`, or `… refusing to bill it through openrouter` |
+| Dev helper | `scripts/harness-run.mjs` | A braked Pi provider, a Pi provider outside the allowlist, a Pi writer other than `openrouter/qwen/qwen3.7-plus`, or an aggregator id whose vendor segment is a native vendor | `pi brake: xai has a native harness; refusing Pi impersonation`, or `… refusing to bill it through openrouter` |
 | Developer roster | `scripts/roster-policy.mjs` | An aggregator route whose vendor segment is a native vendor | `Roster policy refused: native vendor cannot use Pi` |
 
 The dev helper's allowlist (`openrouter`, `nous-portal`, `antigravity`) is a different list from `PI_ALLOWED_PROVIDERS` in `plugins/kxm/src/harness.ts`, which gates nothing. Which of the two should govern is an open decision (Tracking → Still open). Edit permission exists only in the dev helper, and only for admitted writer or experiment routes. The helper accepts only a ChatGPT login for codex, not an API key.
@@ -129,7 +129,7 @@ These extend the generic examples on the reference page with this checkout's adm
 | Billing | grok.com subscription (OAuth) | $2.00 input, $6.00 output, $0.50 cached input |
 | Context | 500,000 (`x-ai/grok-4.7` in the refreshed inventory) | 500,000 |
 
-Pi's own `xai` provider reported `ready` (OAuth) on the capture machine. When the grok quota runs out, the next writer in the lineup is `pi-qwen3-coder-plus-openrouter`, a different vendor.
+Pi's own `xai` provider reported `ready` (OAuth) on the capture machine. When the grok quota runs out, the next writer in the lineup is `pi-qwen3-7-plus-openrouter`, a different vendor.
 
 ### GPT-6.1 Sol
 
@@ -177,11 +177,11 @@ The code differs from that decision: `.kxm/routes.yaml` admits `google/gemini-3.
 | Model | Vendor-plan route | OpenRouter route | Notes |
 |---|---|---|---|
 | Qwen3.8 Flash | `qwen-token-plan/qwen3.8-flash`: removed from the admitted list on 2026-09-27; `qwen-token-plan` is not allowlisted | `openrouter/qwen/qwen3.8-flash`: admitted, route `pi-qwen3-8-flash-openrouter`, read-only fallback for planner and reviewer-cli | `prices.yaml` keeps both rows. |
-| Qwen3 Coder Plus | none | `openrouter/qwen/qwen3-coder-plus`: route `pi-qwen3-coder-plus-openrouter` | The only admitted Pi writer: exact model, `edit` permission. |
+| Qwen3.7 Plus | none | `openrouter/qwen/qwen3.7-plus`: route `pi-qwen3-7-plus-openrouter`, writer #2 | The only admitted Pi writer: exact model, `edit` permission. OpenRouter list on 2026-10-17 is $0.32 input and $1.28 output per 1M tokens, 1,000,000 context, tools supported. |
 | GLM 5.3 and 5.3 Flash | `zai-coding-cn/glm-5.3` and `…/glm-5.3-flash`: removed from the admitted list on 2026-09-27; `zai-coding-cn` is not allowlisted | `openrouter/z-ai/glm-5.3-flash`: admitted, route `pi-glm-5-3-flash-openrouter`, read-only fallback for reviewer-arch | Z.ai has no native harness. |
 | DeepSeek V4.1 Flash | `qwen-token-plan/deepseek-v4.1-flash`: removed from the admitted list on 2026-09-27 | `deepseek/deepseek-v4.1-flash` in the OpenRouter feed | Bills a DeepSeek model through Alibaba's plan. `deepseek` is a native-vendor Pi brake, so this is not a route. |
 
-The developer runner is stricter. Its only Pi writer is `openrouter/qwen/qwen3-coder-plus`, and it does not allowlist `qwen-token-plan` or `zai-coding-cn`. Today OpenRouter is the right answer here for Qwen3 Coder Plus, Qwen3.8 Flash and GLM 5.3 Flash.
+The developer runner is stricter. Its only Pi writer is `openrouter/qwen/qwen3.7-plus`, and it does not allowlist `qwen-token-plan` or `zai-coding-cn`. Today OpenRouter is the right answer here for Qwen3.7 Plus, Qwen3.8 Flash and GLM 5.3 Flash. `qwen/qwen3-coder-plus` is retired.
 
 ### Nous
 

package/docs/glossary.md

@@ -427,7 +427,7 @@ A list of allowed models. It means either a [role](#role)'s roster, or the devel
 
 ### Route
 
-A model selector, such as `xai/grok-4.6` or `openrouter/qwen/qwen3-coder-plus`, that a step can run on. The harness that runs it comes from the agent or the project default. See [Harness routing](reference/harness-routing.md).
+A model selector, such as `xai/grok-4.6` or `openrouter/qwen/qwen3.7-plus`, that a step can run on. The harness that runs it comes from the agent or the project default. See [Harness routing](reference/harness-routing.md).
 
 ### Session
 

package/docs/guides/governed-skills.md

@@ -61,7 +61,7 @@ Write the procedure as a Markdown file. KXM adds `name` and `description` front
 ```bash
 kxm skills create --file flaky.md --name "Triage flaky test" \
   --description "Isolate and classify a flaky test before changing code" \
-  --created-by agent-a --harness pi --models openrouter/qwen/qwen3-coder-plus \
+  --created-by agent-a --harness pi --models openrouter/qwen/qwen3.7-plus \
   --journal journal_abc
 ```
 

package/docs/guides/pi-workers.md

@@ -82,7 +82,7 @@ When no unused fallback remains, the worker waits `KXM_WORKER_PROVIDER_RETRY_MS`
 > [!IMPORTANT]
 > A worker refuses to start on a model whose vendor has its own native harness. Before Pi starts, the worker runs `--model` and every `--fallback-models` entry through the Pi native-vendor brake, and exits 1 with `pi_native_impersonation_blocked` on the first one it refuses: a native provider (`xai/…`), that vendor's own Pi provider (`openai-codex/…`, `kimi-coding/…`), or an aggregator path to it (`openrouter/x-ai/…`). `--dry-run` does not run this check.
 
-Pick every model with [Harness routing](../reference/harness-routing.md), for example `openrouter/qwen/qwen3-coder-plus` with the fallback `openrouter/z-ai/glm-5.3-flash`. The brake cannot check a worker started without `--model`, which runs Pi's default model, or a bare model id that lets Pi choose the provider, so name the provider.
+Pick every model with [Harness routing](../reference/harness-routing.md), for example `openrouter/qwen/qwen3.7-plus` with the fallback `openrouter/z-ai/glm-5.3-flash`. The brake cannot check a worker started without `--model`, which runs Pi's default model, or a bare model id that lets Pi choose the provider, so name the provider.
 
 An agent name is only a label. To use a native-harness model as a peer, connect that harness to the hub under the agent name instead, for example the Claude Code plugin with the name `reviewer-claude`.
 
@@ -203,7 +203,7 @@ kxm hub stop
 | `KXM worker <project>/<name> is already managed by PID <pid>` | A supervisor for this agent is already running. | Use the running worker, or stop it first. |
 | The worker exits with `worker requires --name and --project` | Name or project missing. | Pass `--name` and `--project`, or set `KXM_AGENT_NAME` and `KXM_PROJECT`. |
 | `KXM_WORKER_FALLBACK_MODELS requires KXM_WORKER_MODEL` | Fallbacks without a primary model. | Add `--model`. |
-| The worker exits 1 with `pi_native_impersonation_blocked: …` | `--model` or a fallback is a model whose vendor has its own harness. | Run that model in its native harness, or use an admitted Pi route such as `openrouter/qwen/qwen3-coder-plus`. |
+| The worker exits 1 with `pi_native_impersonation_blocked: …` | `--model` or a fallback is a model whose vendor has its own harness. | Run that model in its native harness, or use an admitted Pi route such as `openrouter/qwen/qwen3.7-plus`. |
 | `kxm connection failed` in the Pi log, or Pi shows `hub:off` | Wrong URL, token, or project, or the name is live elsewhere (`duplicate_agent_name`). | Compare `KXM_SERVER_URL`, `KXM_AUTH_TOKEN`, and `KXM_PROJECT` with the hub; check `kxm peer list`. |
 | Workflow runs share one conversation | Isolation is `off`, the default. | Restart with `--session-isolation workflow`. |
 | A request stays `delivered` | A turn, tool, or provider call is still running, or the watchdogs are recovering it. | Read the lifecycle log. Do not send a duplicate. |

package/docs/guides/provenance-gates.md

@@ -173,7 +173,7 @@ kxm agent worker --name reviewer-grok --project provenance-demo --model <reviewe
 </details>
 
 > [!IMPORTANT]
-> The reviewer names are labels. Choose every Pi model with [Harness routing](../reference/harness-routing.md), for example `openrouter/qwen/qwen3-coder-plus` and `openrouter/z-ai/glm-5.3-flash` for two reviewers from different vendors. A worker refuses a model whose vendor has its own native harness, such as Anthropic or xAI, with `pi_native_impersonation_blocked`. To use such a model as a reviewer, connect its native harness under the same agent name instead, for example the Claude Code plugin as `reviewer-claude`.
+> The reviewer names are labels. Choose every Pi model with [Harness routing](../reference/harness-routing.md), for example `openrouter/qwen/qwen3.7-plus` and `openrouter/z-ai/glm-5.3-flash` for two reviewers from different vendors. A worker refuses a model whose vendor has its own native harness, such as Anthropic or xAI, with `pi_native_impersonation_blocked`. To use such a model as a reviewer, connect its native harness under the same agent name instead, for example the Claude Code plugin as `reviewer-claude`.
 
 In an operator terminal, supply the workflow-start secret, validate the definition, and create a run:
 

package/docs/operations/troubleshooting.md

@@ -134,7 +134,7 @@ The [plugin troubleshooting guide](../../plugins/kxm/README.md#troubleshooting)
 | The heartbeat is healthy but one tool never finishes | A tool exceeded `KXM_WORKER_TOOL_TIMEOUT_MS` (31 minutes by default, above the 30-minute fanout wait) | The worker logs `worker_tool_timeout` and restarts the child; raise the limit only above the longest legitimate call |
 | A long-lived worker keeps restarting | Pi is missing from the service `PATH`, the working directory is gone, or model credentials are missing | Read `worker_process_error` and `worker_exited`; set `KXM_PI_COMMAND` to an explicit path |
 | A read-only reviewer edits files | Prompt wording does not remove tools | Set `KXM_WORKER_TOOLS=read,grep,find,ls` |
-| The worker exits at once with `pi_native_impersonation_blocked` | `--model` or a `--fallback-models` entry is a model whose vendor has its own harness, such as `xai/…`, `openai-codex/…` or `openrouter/x-ai/…` | Run that model in its native harness, or pick an admitted Pi route such as `openrouter/qwen/qwen3-coder-plus`; see [Harness routing](../reference/harness-routing.md#what-the-brake-refuses) |
+| The worker exits at once with `pi_native_impersonation_blocked` | `--model` or a `--fallback-models` entry is a model whose vendor has its own harness, such as `xai/…`, `openai-codex/…` or `openrouter/x-ai/…` | Run that model in its native harness, or pick an admitted Pi route such as `openrouter/qwen/qwen3.7-plus`; see [Harness routing](../reference/harness-routing.md#what-the-brake-refuses) |
 
 Use `--fresh-start`, not `--no-continue`, when only the first launch must avoid old session state.
 

package/docs/reference/cli-reference.md

@@ -1157,7 +1157,7 @@ admitted google/gemini-4-argon-high
 admitted google/gemini-4-argon-medium
 admitted openai/gpt-5.6-sol
 admitted openai/gpt-6.1-sol
-admitted openrouter/qwen/qwen3-coder-plus
+admitted openrouter/qwen/qwen3.7-plus
 admitted openrouter/qwen/qwen3.8-flash
 admitted openrouter/z-ai/glm-5.3-flash
 admitted xai/grok-4.7
@@ -1169,7 +1169,7 @@ reviewer-cli codex-gpt-6-1-sol
 reviewer-cli pi-qwen3-8-flash-openrouter
 reviewer-cli codex-gpt-5-6-sol
 writer grok-grok-4-7
-writer pi-qwen3-coder-plus-openrouter
+writer pi-qwen3-7-plus-openrouter
 writer agy-gemini-4-argon-high
 writer agy-gemini-4-argon-medium
 writer agy-gemini-3-8-flash-medium
@@ -1181,7 +1181,7 @@ kxm routes list --json
 ```
 
 ```text
-{"schema":"kxm.cli-result.v1","ok":true,"command":"routes list","policy":{"schema":"kxm.routes.v2","updatedAt":"2026-10-17T14:30:00.000Z","admitted":["anthropic/fable","google/gemini-3.8-flash-high","google/gemini-3.8-flash-medium","google/gemini-4-argon-high","google/gemini-4-argon-medium","openai/gpt-5.6-sol","openai/gpt-6.1-sol","openrouter/qwen/qwen3-coder-plus","openrouter/qwen/qwen3.8-flash","openrouter/z-ai/glm-5.3-flash","xai/grok-4.7"],"disabled":[]},"membership":["planner claude-fable","planner pi-qwen3-8-flash-openrouter","reviewer-arch claude-fable","reviewer-arch pi-glm-5-3-flash-openrouter","reviewer-cli codex-gpt-6-1-sol","reviewer-cli pi-qwen3-8-flash-openrouter","reviewer-cli codex-gpt-5-6-sol","writer grok-grok-4-7","writer pi-qwen3-coder-plus-openrouter","writer agy-gemini-4-argon-high","writer agy-gemini-4-argon-medium","writer agy-gemini-3-8-flash-medium","writer agy-gemini-3-8-flash-high"]}
+{"schema":"kxm.cli-result.v1","ok":true,"command":"routes list","policy":{"schema":"kxm.routes.v2","updatedAt":"2026-10-17T14:30:00.000Z","admitted":["anthropic/fable","google/gemini-3.8-flash-high","google/gemini-3.8-flash-medium","google/gemini-4-argon-high","google/gemini-4-argon-medium","openai/gpt-5.6-sol","openai/gpt-6.1-sol","openrouter/qwen/qwen3.7-plus","openrouter/qwen/qwen3.8-flash","openrouter/z-ai/glm-5.3-flash","xai/grok-4.7"],"disabled":[]},"membership":["planner claude-fable","planner pi-qwen3-8-flash-openrouter","reviewer-arch claude-fable","reviewer-arch pi-glm-5-3-flash-openrouter","reviewer-cli codex-gpt-6-1-sol","reviewer-cli pi-qwen3-8-flash-openrouter","reviewer-cli codex-gpt-5-6-sol","writer grok-grok-4-7","writer pi-qwen3-7-plus-openrouter","writer agy-gemini-4-argon-high","writer agy-gemini-4-argon-medium","writer agy-gemini-3-8-flash-medium","writer agy-gemini-3-8-flash-high"]}
 ```
 
 ### `kxm routes count`
@@ -1386,7 +1386,7 @@ kxm role get writer --json
 ```
 
 ```text
-{"schema":"kxm.cli-result.v1","ok":true,"command":"role get","roleId":"writer","scope":"local","filePath":"/work/kxm/.kxm/roles/writer.yaml","role":{"schema":"kxm.role.v2","id":"writer","purpose":"writer","permission":"edit","description":"Primary implementation agent.","skills":[],"roster":[{"route":"grok-grok-4-7","effort":"medium"},{"route":"pi-qwen3-coder-plus-openrouter","effort":"medium"},{"route":"agy-gemini-4-argon-high","effort":"medium"},{"route":"agy-gemini-4-argon-medium","effort":"medium"},{"route":"agy-gemini-3-8-flash-medium","effort":"medium"},{"route":"agy-gemini-3-8-flash-high","effort":"medium"}]}}
+{"schema":"kxm.cli-result.v1","ok":true,"command":"role get","roleId":"writer","scope":"local","filePath":"/work/kxm/.kxm/roles/writer.yaml","role":{"schema":"kxm.role.v2","id":"writer","purpose":"writer","permission":"edit","description":"Primary implementation agent.","skills":[],"roster":[{"route":"grok-grok-4-7","effort":"medium"},{"route":"pi-qwen3-7-plus-openrouter","effort":"medium"},{"route":"agy-gemini-4-argon-high","effort":"medium"},{"route":"agy-gemini-4-argon-medium","effort":"medium"},{"route":"agy-gemini-3-8-flash-medium","effort":"medium"},{"route":"agy-gemini-3-8-flash-high","effort":"medium"}]}}
 ```
 
 Captured from this checkout with `node scripts/kxm.mjs role get writer --json`. The path is shortened to `/work/kxm`.
@@ -1396,7 +1396,7 @@ kxm role modify writer --add-skill kxm --dry-run --json
 ```
 
 ```text
-{"schema":"kxm.cli-result.v1","ok":true,"command":"role modify","roleId":"writer","id":"writer","role":{"schema":"kxm.role.v2","id":"writer","purpose":"writer","permission":"edit","description":"Primary implementation agent.","skills":["kxm"],"roster":[{"route":"grok-grok-4-7","effort":"medium"},{"route":"pi-qwen3-coder-plus-openrouter","effort":"medium"},{"route":"agy-gemini-4-argon-high","effort":"medium"},{"route":"agy-gemini-4-argon-medium","effort":"medium"},{"route":"agy-gemini-3-8-flash-medium","effort":"medium"},{"route":"agy-gemini-3-8-flash-high","effort":"medium"}]},"filePath":"/work/kxm/.kxm/roles/writer.yaml","scope":"local","dryRun":true,"planned":[{"action":"write","target":"/work/kxm/.kxm/roles/writer.yaml"}]}
+{"schema":"kxm.cli-result.v1","ok":true,"command":"role modify","roleId":"writer","id":"writer","role":{"schema":"kxm.role.v2","id":"writer","purpose":"writer","permission":"edit","description":"Primary implementation agent.","skills":["kxm"],"roster":[{"route":"grok-grok-4-7","effort":"medium"},{"route":"pi-qwen3-7-plus-openrouter","effort":"medium"},{"route":"agy-gemini-4-argon-high","effort":"medium"},{"route":"agy-gemini-4-argon-medium","effort":"medium"},{"route":"agy-gemini-3-8-flash-medium","effort":"medium"},{"route":"agy-gemini-3-8-flash-high","effort":"medium"}]},"filePath":"/work/kxm/.kxm/roles/writer.yaml","scope":"local","dryRun":true,"planned":[{"action":"write","target":"/work/kxm/.kxm/roles/writer.yaml"}]}
 ```
 
 Add an existing route to a role's roster (Not run):
@@ -2014,11 +2014,11 @@ Starts a long-lived, supervised Pi RPC worker in the foreground through `scripts
 - Needs Pi and a reachable hub. Runs until stopped (`kxm hub stop` stops managed workers too). The exit code is the worker's.
 - A name and project are required (exit 2 otherwise); an invalid isolation mode exits 2.
 - `--dry-run` prints a `kxm.worker-result.v1` envelope with `workspace`, `name`, `project`, `model`, `fallbackModels`, `tools`, `sessionIsolation`, `continue`, `freshStart`.
-- Before Pi starts, the worker runs `--model` and every `--fallback-models` entry through the Pi native-vendor brake. A model whose vendor has its own harness (for example `xai/…`, `openai-codex/…`, `openrouter/x-ai/…` or `antigravity/claude-…`) exits 1 with `pi_native_impersonation_blocked: <message>` on stderr; use the native harness, or an admitted Pi route such as `openrouter/qwen/qwen3-coder-plus`. `--dry-run` does not run this check. See [Harness routing](harness-routing.md#what-the-brake-refuses).
+- Before Pi starts, the worker runs `--model` and every `--fallback-models` entry through the Pi native-vendor brake. A model whose vendor has its own harness (for example `xai/…`, `openai-codex/…`, `openrouter/x-ai/…` or `antigravity/claude-…`) exits 1 with `pi_native_impersonation_blocked: <message>` on stderr; use the native harness, or an admitted Pi route such as `openrouter/qwen/qwen3.7-plus`. `--dry-run` does not run this check. See [Harness routing](harness-routing.md#what-the-brake-refuses).
 - The remaining worker variables are described in [Long-lived worker settings](configuration.md#long-lived-worker-settings).
 
 ```bash
-kxm agent worker --name reviewer --project demo --model openrouter/qwen/qwen3-coder-plus --tools read,grep,find,ls --session-isolation workflow --fresh-start --dry-run
+kxm agent worker --name reviewer --project demo --model openrouter/qwen/qwen3.7-plus --tools read,grep,find,ls --session-isolation workflow --fresh-start --dry-run
 ```
 
 ```text
@@ -2026,7 +2026,7 @@ would start worker
 ```
 
 ```bash
-kxm agent worker --name coordinator --project demo --model openrouter/qwen/qwen3-coder-plus --session-isolation workflow
+kxm agent worker --name coordinator --project demo --model openrouter/qwen/qwen3.7-plus --session-isolation workflow
 ```
 
 Not run: starts a long-lived Pi worker.
@@ -3362,7 +3362,7 @@ Submit a skill candidate from verified episodes.
 - Writes a candidate. Honors `--dry-run`. JSON keys: `metadata` (or `name` for a dry run).
 
 ```bash
-kxm skills create --file SKILL.md --name retry-backoff --created-by alice --harness pi --models openrouter/qwen/qwen3-coder-plus --run wf_123 --dry-run --json
+kxm skills create --file SKILL.md --name retry-backoff --created-by alice --harness pi --models openrouter/qwen/qwen3.7-plus --run wf_123 --dry-run --json
 ```
 
 ```text
@@ -3627,7 +3627,7 @@ Dedicated offline benchmark for side-by-side model comparison (Decision Q12). To
 | Option | Argument | Default | Description |
 |---|---|---|---|
 | `--task` | `<fixture>` | `Deterministic benchmark task` | Task prompt or fixture path for benchmark comparison |
-| `--arms` | `<models>` | `grok/grok-4.6,claude/fable,pi/qwen3-coder-plus` | Comma-separated model routes to benchmark (e.g. grok/grok-4.6,claude/fable) |
+| `--arms` | `<models>` | `grok/grok-4.6,claude/fable,pi/qwen3.7-plus` | Comma-separated model routes to benchmark (e.g. grok/grok-4.6,claude/fable) |
 | `--runs` | `<count>` | `1` | Benchmark runs per arm |
 
 - Reads nothing. JSON keys: `task`, `runs`, `timestamp`, `arms` (`harness`, `model`, `latencyMs`, `tokensIn`, `tokensOut`, `costUsd`, `outcome`).
@@ -3641,7 +3641,7 @@ Routing Benchmark Results (task: Deterministic benchmark task, runs: 1)
 Harness    Model                     Latency(ms)   TokensIn  TokensOut    Cost($)    Outcome
 grok       grok-4.6                          420       1200        450      $0.17     passed
 claude     fable                             680       1200        450      $0.45     passed
-pi         qwen3-coder-plus                  560       1200        450      $0.12     passed
+pi         qwen3.7-plus                  560       1200        450      $0.12     passed
 ```
 
 ## `kxm prices`

package/docs/reference/config-reference.md

@@ -963,7 +963,7 @@ description: Primary implementation agent.
 roster:
   - route: grok-grok-4-7
     effort: medium
-  - route: pi-qwen3-coder-plus-openrouter
+  - route: pi-qwen3-7-plus-openrouter
     effort: medium
 ```
 
@@ -991,7 +991,7 @@ unknown keys are ignored).
 
 A route string is derived from the selected model file's `vendor` and
 `model`, joined by a slash: `xai/grok-4.6`, `anthropic/fable`,
-`openrouter/qwen/qwen3-coder-plus`. When the file is missing, nothing is
+`openrouter/qwen/qwen3.7-plus`. When the file is missing, nothing is
 admitted and every live attempt is refused (`producer_route_unsupported`, or
 `producer_route_not_admitted` from the live producer). A leftover
 `.kxm/producers.yaml` makes every reader fail with `retired
@@ -1002,7 +1002,7 @@ admitted and every live attempt is refused (`producer_route_unsupported`, or
 `model_selection_required`. Inventory IDs come in several shapes (bare
 `grok-4.6` from `grok models`, `vendor/model` from OpenRouter,
 `provider/model` from `pi --list-models`), so a route such as
-`openrouter/qwen/qwen3-coder-plus` or `anthropic/fable` usually has to be
+`openrouter/qwen/qwen3.7-plus` or `anthropic/fable` usually has to be
 added by editing the file.
 
 ```yaml
@@ -1013,7 +1013,7 @@ admitted:
   - openai/gpt-5.6-sol
   - openai/gpt-6.1-sol
   - xai/grok-4.6
-  - openrouter/qwen/qwen3-coder-plus
+  - openrouter/qwen/qwen3.7-plus
 disabled: []
 ```
 
@@ -1068,9 +1068,9 @@ routes:
     roles: [writer]
     permissions: [edit]
     status: admitted            # admitted | retired
-  pi-qwen3-coder-plus-openrouter:
+  pi-qwen3-7-plus-openrouter:
     harness: pi
-    model: openrouter/qwen/qwen3-coder-plus   # Pi: allowed provider prefix + vendor/model
+    model: openrouter/qwen/qwen3.7-plus   # Pi: allowed provider prefix + vendor/model
     vendor: alibaba
     roles: [writer]
     permissions: [edit]
@@ -1097,7 +1097,7 @@ routes:
     permissions: [read-only]
     status: admitted
 lineup:                         # routes admitted for each role, preference order
-  writer: [grok-grok-4-7, pi-qwen3-coder-plus-openrouter, agy-gemini-4-argon-high, agy-gemini-4-argon-medium, agy-gemini-3-8-flash-medium, agy-gemini-3-8-flash-high]
+  writer: [grok-grok-4-7, pi-qwen3-7-plus-openrouter, agy-gemini-4-argon-high, agy-gemini-4-argon-medium, agy-gemini-3-8-flash-medium, agy-gemini-3-8-flash-high]
   planner: [claude-fable, pi-qwen3-8-flash-openrouter]
   reviewer-arch: [claude-fable, pi-glm-5-3-flash-openrouter]
   reviewer-cli: [codex-gpt-6-1-sol, pi-qwen3-8-flash-openrouter, codex-gpt-5-6-sol]
@@ -1105,11 +1105,11 @@ required_critics:
   review-arch: claude-fable
   review-cli: codex-gpt-6-1-sol
 model_origins:                  # required for every Pi route model
-  openrouter/qwen/qwen3-coder-plus:
+  openrouter/qwen/qwen3.7-plus:
     vendor: alibaba
     evidence:
-      source: plans/evidence/route-qwen-openrouter-pi.md
-      sha256: b15f8a54adc398b33e75609e959076eee310bf4ff6623d0b8470ea3da95e6c54
+      source: plans/evidence/route-pi-qwen3-7-plus-openrouter.md
+      sha256: 10e11cfa572ef4891e514d229fa8b7f0f355b31e6481315a6dd763bbe44244a1
 ```
 
 Validated with `validateRosterDocument` from `scripts/roster-policy.mjs`,
@@ -1264,10 +1264,44 @@ file even when a source fails, and exits 1 if any source failed.
 | `models[].sources` | Which sources listed the model |
 | `models[].standard` | OpenRouter per-million rates: `inputPerMillion`, `outputPerMillion`, `cacheReadPerMillion`, `cacheWritePerMillion` |
 | `models[].discount` | Nous per-million rates, same fields |
+| `models[].missingUpstream` | `true` when the weekly refresh kept a row that OpenRouter no longer lists because a route or roster still names it |
 
 The inventory is discovery data, not admission. Admitting a route is a
 separate, reviewed change to `.kxm/routes.yaml`.
 
+### Weekly OpenRouter refresh
+
+`.github/workflows/refresh-models.yml` runs every Monday at 13:00 UTC and on
+`workflow_dispatch`. It fetches `https://openrouter.ai/api/v1/models` and
+updates the OpenRouter `standard` rates in this inventory, plus the matching
+single-tier rows in `.kxm/prices.yaml`. Nous `discount` rates and the `pi`,
+`grok`, and `agy` source blocks stay as the last `kxm models inventory-refresh`
+left them. Output is sorted by model id. `prices.yaml` `date` and the
+inventory `fetchedAt` (midnight UTC of that date) change only when the catalog
+content changes, so an unchanged week does not open a pull request.
+
+A model named by `.kxm/routes.yaml` (`admitted` or `disabled`) or by a roster
+entry in `.kxm/roles/*.yaml` is not deleted when it leaves OpenRouter. The
+refresh keeps the previous inventory row, sets `missingUpstream: true`, and
+the workflow's `model-catalog-guard` check fails. The pull request lists that
+model, and any referenced model whose input or output rate moved by more than
+20 percent, at the top. A multi-tier price row is left unchanged and called
+out the same way. `anthropic/fable` is matched through its aliases; when those
+OpenRouter rows disagree, the price row is left unchanged.
+
+Run the same refresh from a checkout:
+
+```bash
+node --disable-warning=ExperimentalWarning --experimental-strip-types scripts/refresh-model-catalog.mjs --dry-run
+node --disable-warning=ExperimentalWarning --experimental-strip-types scripts/refresh-model-catalog.mjs
+```
+
+`--dry-run` prints the summary and writes nothing. Exit 2 means a referenced
+model would be kept after leaving the catalog. The workflow pushes a non-empty
+diff to `automation/model-refresh` and opens or updates one pull request with
+`GITHUB_TOKEN`. It then dispatches `ci.yml` on that branch, because a push
+made with `GITHUB_TOKEN` does not start CI by itself.
+
 ## Personalization settings (`kxm.config.v1`)
 
 Personal and workflow preferences, merged in three layers: built-in defaults

package/docs/reference/configuration.md

@@ -166,14 +166,14 @@ For example, a review-only worker that cannot edit the checkout:
 
 ```bash
 export KXM_WORKER_TOOLS="read,grep,find,ls,kxm_list,kxm_get,kxm_reply"
-kxm agent worker --name reviewer --project <project> --model openrouter/qwen/qwen3-coder-plus
+kxm agent worker --name reviewer --project <project> --model openrouter/qwen/qwen3.7-plus
 ```
 
 <details><summary>PowerShell</summary>
 
 ```powershell
 $env:KXM_WORKER_TOOLS = "read,grep,find,ls,kxm_list,kxm_get,kxm_reply"
-kxm agent worker --name reviewer --project <project> --model openrouter/qwen/qwen3-coder-plus
+kxm agent worker --name reviewer --project <project> --model openrouter/qwen/qwen3.7-plus
 ```
 
 </details>