workspai 0.78.0 → 0.79.1

package/docs/README.md

@@ -1,109 +1,65 @@
-# Workspai NPM — Documentation Index
+# Workspai documentation
 
-Workspai helps people and tools understand the same software system. Use these
-guides to connect existing projects, create new ones, map their relationships,
-check changes, prepare AI context, and automate release checks.
+**New here? Start with your project.** You do not need to learn the architecture
+or move your source to try Workspai.
 
-Start with the [main README](../README.md) for the product overview, or use the
-quickstart below. `workspai` is the main package and command; `wspai` is only a
-shorter optional name.
+[Product overview](../README.md) · [Command reference](./commands-reference.md) ·
+[Glossary](./GLOSSARY.md)
 
-## Quickstart: from zero to value in 60 seconds
+## First run
 
-You have an existing project. You want AI agents (and your team) to understand
-it as a system—not just a pile of files. Three commands:
+You need **Node.js 20.19+ and npm**. In your project's terminal:
 
 ```bash
-# Step 1: Adopt the project (it stays where it is)
-cd /path/to/your/project
 npx workspai adopt .
+npx workspai workspace intelligence run --for-agent generic
+```
 
-# Step 2: Build the full workspace intelligence
-npx workspai workspace intelligence run --for-agent generic --strict --json
+Workspai adds metadata and agent entry files to your project and saves shared
+reports in a linked workspace. The system map and checks do not require an AI
+API key. Read the terminal summary to find context and checks that need attention.
 
-# Step 3: Validate the canonical agent entry
-npx workspai agent bootstrap --for-agent generic --strict --json
-```
+Starting a new project? Run `npx workspai create` for guided setup.
 
-That is it. Your project now has:
+## What do you want to do next?
 
-- A **Workspace Model** describing every project, runtime, and dependency
-- A **Knowledge Graph** with searchable, proof-backed relationships
-- **Agent grounding** files (`AGENTS.md`, Cursor rules, Claude rules, etc.)
-- **Health checks** that name what is broken and why
+| Your goal | Start here |
+| --- | --- |
+| Set up an existing or new project | [Getting started](./creating-workspaces-and-projects.md) |
+| Connect Codex, Claude Code, Cursor, or Copilot | [Agent setup](./agent-entry.md) |
+| Find relationships and their source references | [Explore the graph](./workspace-knowledge-graph.md) |
+| Give an agent a specific outcome | [Plan a goal](./goal-packs.md) |
+| Record and verify an agent's changes | [Proof-Carrying Change](./proof-carrying-change.md) |
+| Fix a reported blocker | [Doctor](./doctor-command.md) · [Repair workflow](./workspace-repair-engine.md) |
+| Run checks in CI | [CI guide](./ci-workflows.md) |
 
-Continue from the same project terminal for any Workspai command—no `cd` needed.
+You can keep working from the same project terminal. After changes, rerun the
+intelligence command to refresh the saved view. A blocked check names something
+to investigate; an empty graph search means no current evidence proves a match.
 
-Use `generic` for a vendor-neutral context pack, or select a published host
-target such as `codex`, `claude`, `gemini`, `qwen`, `kimi`, `grok`, `copilot`,
-`cursor`, `windsurf`, or `amazon-q`. Agent Sync publishes shared files and thin
-host adapters that route every consumer to the same evidence.
+## For CI and automation
 
-Use the release pipeline when you need the broader release workflow:
+Use the strict JSON runner:
 
 ```bash
-npx workspai pipeline --json --strict
+npx workspai workspace intelligence run --for-agent generic --strict --json
 ```
 
-Adoption keeps source in place. The command maps the system and its
-connections, checks what changes may affect, runs health and release checks,
-and prepares focused AI context under `.workspai/`. When the workspace is not
-ready, the report shows the reason and points to the files or reports behind
-it. Automation details and supported AI tools are documented in the
-[Unified runner](./workspace-intelligence-runner.md). See the
-[Artifact Catalog](./contracts/ARTIFACT_CATALOG.md) for exact paths, writers,
-schemas, and consumers.
+[Runner behavior and exit codes](./workspace-intelligence-runner.md) ·
+[Generated files and schemas](./contracts/ARTIFACT_CATALOG.md)
 
-To start with a new workspace instead:
+## Building agents or model gateways?
 
-```bash
-npx workspai create workspace my-workspace --profile minimal --yes
-cd ~/.workspai/workspaces/my-workspace
-npx workspai create project nextjs web --yes
-```
+[Agent framework adapters](./agent-framework-adapters.md) covers the supported
+framework kits. [Model gateways](./model-gateways.md) covers server-owned SDK
+starters. These are separate from giving a coding agent context about your project.
 
-From the workspace terminal, create a project, use `adopt` to link one in place,
-or use `import` to copy or clone one into the workspace.
-
-## Table of contents
-
-- [Choose a guide by goal](#choose-a-guide-by-goal)
-- [User documentation](#user-documentation)
-- [Operations & security](#operations--security)
-- [AI module recommendations](#ai-module-recommendations)
-- [Technical contracts](#technical-contracts)
-- [Contributor documentation](#contributor-documentation)
-- [Validation commands](#validation-commands)
-
-## Choose a guide by goal
-
-| I want to…                                         | Start here                                                                                  | Expected outcome                                                         |
-| -------------------------------------------------- | ------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------ |
-| Create a workspace or project                      | [Creating workspaces and projects](./creating-workspaces-and-projects.md)                   | A registered project with canonical `.workspai` metadata                 |
-| Bring an existing repository under governance      | [Workspace operations](./workspace-operations.md#import-and-adoption)                       | Source stays in place with `adopt`, or is copied/cloned with `import`    |
-| Run the complete intelligence loop                 | [Unified runner](./workspace-intelligence-runner.md)                                        | One ordered run report with durable stage evidence                       |
-| Turn plain language into governed work             | [Goal Packs](./goal-packs.md)                                                               | A scope-bound, evidence-pinned plan and portable agent handoff           |
-| Prove what an agent changed and why                | [Proof-Carrying Change](./proof-carrying-change.md)                                         | A Goal-bound, effect-receipted, independently verified change capsule    |
-| Define or implement an agent framework integration | [Agent Framework Adapter Contract](./agent-framework-adapters.md)                           | One provider-neutral manifest, ownership boundary, and conformance gate  |
-| Create a server-owned model gateway                | [AI Gateway](./model-gateways.md)                                                           | OpenRouter TypeScript or Python Client SDK starter with workspace lifecycle |
-| Ground an agent before broad source discovery      | [Canonical-first agent entry](./agent-entry.md)                                             | A portable receipt for host discovery, evidence integrity, and freshness |
-| Repair a blocker through an approved transaction   | [Workspace Repair Engine](./workspace-repair-engine.md)                                     | Checkpointed execution, validation, canonical verify, and safe rollback  |
-| Observe current CLI and Studio activity            | [Workspai Live](./workspace-live-activity.md)                                               | One bounded activity projection for terminal, IDE, replay, and capture   |
-| Set a release, security, or coverage outcome       | [Verified engineering goals](./workspace-intelligence-runner.md#verified-engineering-goals) | A durable success contract with a current evidence-backed verdict        |
-| Ask an architecture or dependency question         | [Workspace Knowledge Graph](./workspace-knowledge-graph.md)                                 | A bounded answer with proof references rather than the whole graph       |
-| Measure agent token, cost, and outcome efficiency  | [Workspace Intelligence Evaluation](./workspace-intelligence-evaluation.md)                 | A live, provenance-aware report suitable for CLI, IDE, and CI            |
-| Benchmark bounded retrieval across fixed scenarios | [Workspace Intelligence Benchmark](./workspace-intelligence-benchmark.md)                   | An offline suite with honest estimate/measurement boundaries             |
-| Integrate CI or release gates                      | [CI workflows](./ci-workflows.md)                                                           | Machine-readable exit codes and uploadable evidence                      |
-| Find the writer, schema, or path for an output     | [Artifact Catalog](./contracts/ARTIFACT_CATALOG.md)                                         | One canonical source instead of path guessing                            |
-| Understand Workspai terminology                    | [Glossary](./GLOSSARY.md)                                                                   | Shared meanings for model, graph, evidence, gate, and artifacts          |
-| Review or change the main product README           | [README content contract](./README_CONTENT_CONTRACT.md)                                     | Stable narrative, claim boundaries, and machine-enforced drift rules     |
-| Choose a first contribution path                   | [Contribution Hub](../../../.github/CONTRIBUTING.md)                                        | One bounded task, support route, and file-aware validation plan          |
-| Contribute to the CLI                              | [Development](./DEVELOPMENT.md)                                                             | Local build, test, contract, and documentation gates                     |
-
-There are two different AI-facing features. Workspace Intelligence is
-deterministic, proof-backed, and does not require an AI API key. The optional
-module recommender uses embeddings to suggest FastAPI or NestJS modules; start
-with [AI Quickstart](./AI_QUICKSTART.md) only when that is your goal.
+## Reference library
+
+Looking for a specific feature, integration, or contract? Open the complete index.
+
+<details>
+<summary>All guides, technical contracts, and contributor resources</summary>
 
 ## User documentation
 
@@ -210,35 +166,11 @@ npm run validate:docs-examples # example JSON/YAML in docs
 npm run smoke:readme           # CLI help smoke for documented commands
 ```
 
-## Repository layout
-
-```text
-workspai/
-├── README.md                 # Monorepo overview
-├── package.json              # Private workspace root
-└── packages/
-    └── cli/
-        ├── README.md         # CLI user hub (install, quickstarts, doc links)
-        ├── CHANGELOG.md
-        ├── RELEASE_NOTES.md
-        ├── releases/         # Per-version release notes
-        └── docs/
-            ├── README.md     # This index
-            ├── README_CONTENT_CONTRACT.md
-            ├── commands-reference.md
-            ├── workspace-knowledge-graph.md
-            ├── workspace-intelligence-evaluation.md
-            ├── workspace-operations.md
-            ├── workspace-run.md
-            ├── ci-workflows.md
-            ├── doctor-command.md
-            ├── OPEN_SOURCE_USER_SCENARIOS.md
-            ├── config-file-guide.md
-            ├── SECURITY.md
-            ├── SETUP.md
-            ├── DEVELOPMENT.md
-            ├── contracts/    # Contract docs (mirrors + matrices)
-            └── …             # AI guides, policies, examples
-```
 
-Enterprise governance runbooks are maintained outside this OSS docs tree.
+</details>
+
+## Help and feedback
+
+[Ask a question](https://github.com/chistiq/workspai/discussions) ·
+[Report a problem](https://github.com/chistiq/workspai/issues) ·
+[Contribute](../../../.github/CONTRIBUTING.md)

package/docs/agent-framework-adapters.md

@@ -100,15 +100,15 @@ verified from the Microsoft source tree:
 
 | Adapter                            | Tested framework | Runtime         | Authored detection                         |
 | ---------------------------------- | ---------------- | --------------- | ------------------------------------------ |
-| `microsoft-agent-framework-python` | `1.18.0`         | Python `>=3.10` | `agent-framework` PyPI package family      |
-| `microsoft-agent-framework-dotnet` | `1.21.0`         | .NET `>=8.0`    | `Microsoft.Agents.AI` NuGet package family |
+| `microsoft-agent-framework-python` | `1.19.0`         | Python `>=3.10` | `agent-framework` PyPI package family      |
+| `microsoft-agent-framework-dotnet` | `1.22.0`         | .NET `>=8.0`    | `Microsoft.Agents.AI` NuGet package family |
 
 The first provider profile pins its integration independently: Python uses
-`agent-framework-foundry` `1.13.0`, while .NET uses
-`Microsoft.Agents.AI.Foundry` `1.21.0-preview.260911.1` over the stable
-`Microsoft.Agents.AI` `1.21.0` core. Preview integration status is not
+`agent-framework-foundry` `1.13.1`, while .NET uses
+`Microsoft.Agents.AI.Foundry` `1.22.0-preview.260918.1` over the stable
+`Microsoft.Agents.AI` `1.22.0` core. Preview integration status is not
 misrepresented as framework stability. The dedicated .NET test project pins
-`Microsoft.NET.Test.Sdk` `18.10.0` and `xunit.v3.mtp-v2` `4.0.1`.
+`Microsoft.NET.Test.Sdk` `18.10.1` and `xunit.v3.mtp-v2` `4.0.1`.
 
 Each adapter implements detection, scaffold and attach planning, managed-file
 rendering, project context, validation, and runtime resolution. It never
@@ -129,7 +129,9 @@ Loaders share the OpenAI containment contract: canonical path walk, regular-file
 open, 128 KiB cap, UTF-8 JSON, and `schemaVersion: project-context-agent.v1`.
 The Python starter is pip-editable (`[build-system]` + setuptools modules) and
 uses the official Foundry hello-world `Agent(client=FoundryChatClient(...))`
-pattern with three read-only documented tools: `describe_workspai_context`,
+pattern. Agent execution enters the SDK agent as an async context manager so
+framework resources are closed according to the current upstream lifecycle.
+The starter exposes three read-only documented tools: `describe_workspai_context`,
 `read_workspai_project_summary`, and `list_workspai_supported_commands`. The
 .NET starter uses `AIProjectClient.AsAIAgent` plus `AIFunctionFactory.Create`,
 enables `RestorePackagesWithLockFile`, and does not emit invented NuGet lock
@@ -168,11 +170,11 @@ not reuse Microsoft detection, kits, or model-provider defaults.
 
 | Adapter                    | Tested framework | Runtime         | Authored detection                 |
 | -------------------------- | ---------------- | --------------- | ---------------------------------- |
-| `openai-agents-python`     | `0.22.2`         | Python `>=3.10` | exact PyPI package `openai-agents` |
+| `openai-agents-python`     | `0.22.3`         | Python `>=3.10` | exact PyPI package `openai-agents` |
 | `openai-agents-typescript` | `0.18.0`         | Node.js `>=22`  | exact npm package `@openai/agents` |
 
-The TypeScript starter also pins peer `zod` `4.6.5`, TypeScript `5.9.3`, and
-`@types/node` `22.20.3`. The `openai` PyPI or npm package alone is not this
+The TypeScript starter also pins peer `zod` `4.6.5`, TypeScript `7.0.2`, and
+`@types/node` `26.6.3`. The `openai` PyPI or npm package alone is not this
 framework. Generated markers under `.workspai/agent-frameworks/` cannot select
 it.
 
@@ -205,7 +207,7 @@ Claimed capabilities are conservative and independently evidenced:
   hosted tools, and human-approval loops
 
 Python cancellation uses `Runner.max_turns` and `ModelSettings.timeout` from
-`openai-agents` `0.22.2`. `ModelSettings.timeout` is a per-model-request
+`openai-agents` `0.22.3`. `ModelSettings.timeout` is a per-model-request
 timeout applied only on the live model path. Injected `ScriptedModel` runs omit
 it because that setting hung the official test double on Python 3.13.
 TypeScript cancellation uses `maxTurns` plus `AbortSignal` from
@@ -233,8 +235,8 @@ workspace profile. A later adapter-manifest change must not hide a kit.
 Create and Attach still refuse a kit whose adapter is not release-admitted;
 visibility in the picker is not permission to write a blocked adapter.
 Attach still requires `--framework openai-agents` when the
-runtime is shared. OpenAI adapters are labeled `stable`; release admission
-remains blocked until the exact v2 cross-platform candidate is promoted.
+runtime is shared. OpenAI adapters are labeled `stable`; Python `0.22.3` and
+TypeScript `0.18.0` are release-admitted from the complete reviewed matrix.
 Handoffs, MCP, sessions, voice, sandbox, and approval loops stay unsupported.
 Microsoft adapters remain `preview`.
 
@@ -250,20 +252,23 @@ Exact pins live in `src/agent-frameworks/version-baselines.v1.json`. Discovery
 must re-prove registry and GitHub agreement on the day it runs; TypeScript 2.0
 graph Workflow Runtime is not generalized to Python.
 
-| Adapter                 | Runtime           | Authored detection            |
-| ----------------------- | ----------------- | ----------------------------- |
+| Adapter                 | Runtime           | Authored detection              |
+| ----------------------- | ----------------- | ------------------------------- |
 | `google-adk-python`     | Python `>=3.10`   | exact PyPI package `google-adk` |
 | `google-adk-typescript` | Node.js `>=20.19` | exact npm package `@google/adk` |
 
-The TypeScript starter also pins `zod`, TypeScript, and `@types/node` from the
-same baseline document. `@google/adk-devtools` is not a v1 dependency. Do not
+The Python baseline is `google-adk` `2.10.0`. The TypeScript baseline is
+`@google/adk` `2.1.0` with `zod` `4.6.5`, TypeScript `7.0.2`, and
+`@types/node` `26.6.3`. `@google/adk-devtools` is not a v1 dependency. Do not
 run unqualified `npx adk`.
 
 Adapters are labeled `preview`. Create kits are `agent.google-adk.python` and
-`agent.google-adk.typescript`. The reviewed v2 inventory admits both exact
-baselines from Linux, macOS, and Windows evidence, so Create and Attach are
-enabled. The generic admission boundary still refuses any unpublished or
-drifted kit before `--dry-run` and before filesystem or workspace mutation.
+`agent.google-adk.typescript`. Python `2.10.0` and TypeScript `2.1.0` are
+admitted by the reviewed v2 Linux, macOS, and Windows evidence. The generic
+admission boundary refuses any unpublished or drifted kit before
+`--dry-run` and before filesystem or workspace mutation. Generated Python code
+uses the current ADK lifecycle: construct `App(root_agent=...)`, pass that app
+to `Runner`, and enter the runner as an async context manager.
 Streaming emits `event.partial` fragments as
 display deltas. The SDK-recognized final response is retained as canonical text
 and is not re-emitted after streamed fragments. Intermediate metadata does not
@@ -342,8 +347,8 @@ npx workspai agent framework plan \
 When more than one published framework shares a runtime, pass `--framework`
 explicitly. Workspai does not guess or fall back. `--runtime python` without
 `--framework` requires an explicit choice because Microsoft Agent Framework,
-OpenAI Agents SDK, and Google ADK are all published and release-admitted at
-their exact reviewed baselines.
+OpenAI Agents SDK, and Google ADK are all published. A published choice may
+still be blocked when its current baseline is not release-admitted.
 
 The interactive attach command displays the same plan and asks before granting
 its filesystem effect. Automation must opt in with `--yes` and records the
@@ -372,7 +377,8 @@ consequences rather than unrelated surprises. A blocked or no-op
 request created without an external Goal closes its generated Change and Goal
 instead of leaving actionable lifecycle state behind.
 
-The same admitted lifecycle is available for a new project:
+The same admission-governed lifecycle is available for a new project when the
+selected baseline is currently admitted:
 
 ```bash
 npx workspai create project agent.microsoft.python support-agent
@@ -454,8 +460,8 @@ multi-language Microsoft Agent Framework. OpenAI Agents SDK and Google ADK
 Python and TypeScript adapters share that same create, attach, detection,
 ownership, and verification host. They are selectable for Create and Attach
 only while their exact manifest digest, framework baseline, runtime, and
-platform list remain in the reviewed release-admission inventory. Microsoft
-adapters remain selectable while their current inventory entries stay valid.
+platform list remain in the reviewed release-admission inventory. All six
+current baselines are admitted by Agent Framework Adapter Matrix run 88.
 
 ## Implementation sequence
 

package/docs/ci-workflows.md

@@ -13,7 +13,7 @@ Map of GitHub Actions workflows in this repository. Use this when editing CI to
 | Official generator smoke  | `.github/workflows/frontend-generator-smoke.yml`          | Contract-driven official-generator drift gate                                             |
 | Agent Framework matrix    | `.github/workflows/agent-framework-conformance.yml`       | Affected-family Linux PR gate plus manual twelve-lane release qualification and admission |
 | Agent Framework discovery | `.github/workflows/agent-framework-version-discovery.yml` | Weekly PyPI/NuGet candidate report; no commit, PR, or contract rewrite                    |
-| Model Gateway matrix      | `.github/workflows/model-gateway-qualification.yml`       | Three-OS generated-project and lifecycle qualification; manual fast mode stays on Linux |
+| Model Gateway matrix      | `.github/workflows/model-gateway-qualification.yml`       | Three-OS generated-project and lifecycle qualification; manual fast mode stays on Linux   |
 | Model Gateway discovery   | `.github/workflows/model-gateway-version-discovery.yml`   | Weekly npm/PyPI/GitHub candidate report; no baseline write, commit, or PR                 |
 | Security                  | `.github/workflows/security.yml`                          | Path-aware scanning plus the always-resolved `Security Gate`                              |
 | Manual npm release        | `.github/workflows/release-npm-manual.yml`                | Maintainer-only release gate and publish workflow                                         |
@@ -68,6 +68,12 @@ from the changed paths:
 | Dependency or security surface         | npm audit, SBOM, dependency review, and CodeQL                         |
 
 Specialized workflows use identical `push` and `pull_request` path filters.
+The complete Windows CLI suite runs in four shards on separate runners, with
+two Vitest workers per runner to limit filesystem contention. Every shard must
+pass before `CI Gate` succeeds; no test files are excluded. Each shard uploads
+a JSON test report with timings. Linux retains the complete coverage suite;
+macOS retains the complete platform suite. Sharding reduces elapsed time by
+using additional runners; it does not promise lower total runner usage.
 Manual and scheduled runs remain available for full qualification and upstream
 drift detection. Workflow definitions and release-gate definitions trigger the
 lanes they govern.
@@ -139,7 +145,7 @@ Validate or preview the current CLI announcement locally:
 npm --workspace workspai run check:release-announcement
 npm --workspace workspai run release:announcement -- \
   --product workspai-cli \
-  --tag v0.78.0 \
+  --tag v0.79.1 \
   --markdown-output /tmp/workspai-discord-announcement.md
 ```
 

package/docs/commands-reference.md

@@ -278,11 +278,12 @@ See [Canonical-first agent entry](./agent-entry.md).
 
 `agent framework` is the governed bridge between Workspai evidence and an
 agent runtime. `list` exposes every built-in adapter and its release-admission
-state. In this CLI version Microsoft Python `1.18.0` and .NET `1.21.0` remain
-`preview`. OpenAI Agents SDK Python `0.22.2` and TypeScript `0.18.0` are
-labeled `stable`. Google ADK Python `2.9.2` and TypeScript `2.1.0` remain
-labeled `preview` and are admitted for Create and Attach by their complete
-Linux, macOS, and Windows evidence. Create and Attach require the reviewed v2 release inventory,
+state. In this CLI version Microsoft Python `1.19.0` and .NET `1.22.0` remain
+`preview`. OpenAI Agents SDK Python `0.22.3` and TypeScript `0.18.0` are
+labeled `stable`. Google ADK Python `2.10.0` and TypeScript `2.1.0` remain
+labeled `preview`. All six current baselines are admitted by the complete
+Linux, macOS, and Windows evidence from Agent Framework Adapter Matrix run 88.
+Create and Attach require the reviewed v2 release inventory,
 whose manifest, framework baseline, runtime, and platform claims were promoted
 from the Linux, macOS, and Windows release matrix. Semantic implementation
 digests remain audit provenance rather than runtime authorization. `plan`
@@ -292,13 +293,13 @@ selects `python`, `dotnet`, or `node`. `--framework` selects the independent
 framework id when more than one admitted adapter shares that runtime.
 `attach` shows that plan and requires an
 interactive confirmation or explicit `--yes` before granting the filesystem
-effect and writing an isolated `agents/<name>` directory. `create project
-agent.microsoft.python|dotnet` uses the same admitted lifecycle for a new
-project: it registers the project, plans against a Model baseline, writes the
-nested runtime, then re-observes Model/Graph before it claims Intelligence is
-sealed. `create project agent.openai.python|typescript` uses the same admitted
-lifecycle. `create project agent.google-adk.python|typescript` uses the same
-admitted lifecycle. Dependency installation, credentials, generated-code execution, and
+effect and writing an isolated `agents/<name>` directory. When its selected
+baseline is admitted, `create project agent.microsoft.python|dotnet` uses the
+same governed lifecycle for a new project: it registers the project, plans
+against a Model baseline, writes the nested runtime, then re-observes
+Model/Graph before it claims Intelligence is sealed. The OpenAI and Google ADK
+Create kits use the same admission-governed lifecycle. Dependency installation,
+credentials, generated-code execution, and
 model provider calls are never implied by that approval. `apply` is the
 automation counterpart for a plan that was separately authorized with
 `change authorize`. Any adapter, version, manifest, runtime, or platform drift

package/docs/create-planner-capabilities.md

@@ -45,7 +45,7 @@ by Workspai and then registered in Workspace Intelligence:
 - Astro: `npm create astro@latest <name>`
 - SvelteKit: `npx sv@latest create <name>`
 - Tauri: `npm create tauri-app@latest <name> -- --template vanilla-ts`
-- Electron Forge: `npx create-electron-app@latest <name> --template=vite-typescript`
+- Electron Forge: `npx create-electron-app@latest <name> --template vite --typescript`
 - VS Code Extension: `npx --package yo@latest --package generator-code@latest -- yo code <name> ...`
 - Laravel: `composer create-project --no-interaction --prefer-dist --stability=stable laravel/laravel <name>`
 

package/docs/creating-workspaces-and-projects.md

@@ -11,6 +11,11 @@ commands, automation, and the most useful options.
 For a compact list of command syntax, see
 [commands-reference.md](./commands-reference.md).
 
+During interactive `workspai create` setup, selection menus offer **← Back**.
+Text fields accept **`/back`** to return to the previous setup step. Backtracking
+changes your choices before creation begins; it does not undo files or navigate
+prompts owned by an external generator. Ctrl+C still cancels the flow.
+
 ## The two things you can create
 
 A **workspace** is the shared home for related projects, rules, and saved
@@ -391,22 +396,24 @@ project metadata and performs the selected workspace registration.
 
 ## Agent Framework kits
 
-| Kit                      | Runtime | Tested baseline | Layout |
-| ------------------------ | ------- | --------------- | ------ |
-| `agent.microsoft.python` | Python  | `agent-framework-core` `1.18.0`, Foundry `1.13.0`, `azure-identity` `1.25.3` | Isolated `agents/<instance>/` with `pyproject.toml`, credentialless `unittest`, and `.env.example` |
-| `agent.microsoft.dotnet` | .NET    | `Microsoft.Agents.AI` `1.21.0`, Foundry `1.21.0-preview.260911.1` | Isolated `agents/<instance>/` with the executable project plus a dedicated test project |
-| `agent.openai.python` | Python | `openai-agents` `0.22.2` | Isolated `agents/<instance>/` with pip-editable `pyproject.toml`, credentialless `unittest`, and `.env.example` |
-| `agent.openai.typescript` | Node.js | `@openai/agents` `0.18.0`, `zod` `4.6.5` | Isolated `agents/<instance>/` with `package.json`, credentialless `node:test`, and `.env.example` |
-| `agent.google-adk.python` | Python | `google-adk` `2.9.2` | Isolated `agents/<instance>/` with pip-editable `pyproject.toml`, credentialless `unittest`, and provider profiles `gemini-api` / `vertex-ai`. Preview and release-admitted. |
-| `agent.google-adk.typescript` | Node.js | `@google/adk` `2.1.0`, `zod` `4.6.5` | Isolated `agents/<instance>/` with `package.json`, credentialless `node:test`, and provider profiles `gemini-api` / `vertex-ai`. Preview, release-admitted, and independent from the Python runtime. |
-
-Interactive `workspai create` shows these kits under **AI Agent** after
-reviewed release admission. Agent kits require Workspace governance and therefore do not
-accept `--no-workspace`. Optional `--agent-name` names the instance directory
-(default `primary`). They do not install dependencies, call a model, or store
-credentials. Exact versions are promoted only after the full Linux, macOS, and
-Windows conformance matrix passes. Weekly registry discovery is report-only; it
-cannot rewrite these pins.
+| Kit                           | Runtime | Tested baseline                                                              | Layout                                                                                                                                                                                               |
+| ----------------------------- | ------- | ---------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| `agent.microsoft.python`      | Python  | `agent-framework-core` `1.19.0`, Foundry `1.13.1`, `azure-identity` `1.25.3` | Isolated `agents/<instance>/` with `pyproject.toml`, credentialless `unittest`, and `.env.example`. Preview and release-admitted.                                                                    |
+| `agent.microsoft.dotnet`      | .NET    | `Microsoft.Agents.AI` `1.22.0`, Foundry `1.22.0-preview.260918.1`            | Isolated `agents/<instance>/` with the executable project plus a dedicated test project. Preview and release-admitted.                                                                               |
+| `agent.openai.python`         | Python  | `openai-agents` `0.22.3`                                                     | Isolated `agents/<instance>/` with pip-editable `pyproject.toml`, credentialless `unittest`, and `.env.example`. Stable and release-admitted.                                                        |
+| `agent.openai.typescript`     | Node.js | `@openai/agents` `0.18.0`, `zod` `4.6.5`                                     | Isolated `agents/<instance>/` with `package.json`, credentialless `node:test`, and `.env.example`                                                                                                    |
+| `agent.google-adk.python`     | Python  | `google-adk` `2.10.0`                                                        | Isolated `agents/<instance>/` with pip-editable `pyproject.toml`, credentialless `unittest`, and provider profiles `gemini-api` / `vertex-ai`. Preview and release-admitted.                         |
+| `agent.google-adk.typescript` | Node.js | `@google/adk` `2.1.0`, `zod` `4.6.5`                                         | Isolated `agents/<instance>/` with `package.json`, credentialless `node:test`, and provider profiles `gemini-api` / `vertex-ai`. Preview, release-admitted, and independent from the Python runtime. |
+
+Interactive `workspai create` shows every published kit under **AI Agent** and
+labels kits whose current baseline awaits release admission. Visibility is not
+authorization: scaffolding remains blocked until the exact baseline has
+reviewed evidence. Agent kits require Workspace governance and therefore do
+not accept `--no-workspace`. Optional `--agent-name` names the instance
+directory (default `primary`). They do not install dependencies, call a model,
+or store credentials. Exact versions are promoted only after the full Linux,
+macOS, and Windows conformance matrix passes. Weekly registry discovery is
+report-only; it cannot rewrite these pins.
 
 Create does not emit an empty root `pyproject.toml` or `.csproj`. The nested
 instance is the only Workspace Run unit. After Create, Model and Graph already
@@ -423,17 +430,17 @@ independent proof for this scaffold. See
 
 ## AI Gateway kits
 
-| Kit | Runtime | Tested baseline | Layout |
-| --- | --- | --- | --- |
-| `gateway.openrouter.typescript` | Node.js | `@openrouter/sdk` `1.3.11`, TypeScript `5.9.3`, `@types/node` `22.20.3` | Server-owned OpenRouter Client SDK starter with `gateway.policy.json`, credentialless `node:test`, and `.env.example` |
-| `gateway.openrouter.python` | Python | `openrouter` `1.2.11` | Server-owned OpenRouter Client SDK starter with `gateway.policy.json`, credentialless `unittest`, and `.env.example` |
+| Kit                             | Runtime | Tested baseline                                                         | Layout                                                                                                                |
+| ------------------------------- | ------- | ----------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------- |
+| `gateway.openrouter.typescript` | Node.js | `@openrouter/sdk` `1.3.33`, TypeScript `5.9.3`, `@types/node` `22.20.3` | Server-owned OpenRouter Client SDK starter with `gateway.policy.json`, credentialless `node:test`, and `.env.example` |
+| `gateway.openrouter.python`     | Python  | `openrouter` `1.2.32`                                                   | Server-owned OpenRouter Client SDK starter with `gateway.policy.json`, credentialless `unittest`, and `.env.example`  |
 
 Interactive `workspai create` shows these kits under **AI Gateway** with the
 hint **Unified model access and routing**. A model identifier is runtime
 configuration, not a kit. Create does not install dependencies, call a model,
 or store credentials. Attach is not supported for Gateway in this release.
-The official Go Client SDK remains beta (`v0.8.11` on
-2026-09-21) and is not admitted. See
+The official Go Client SDK remains beta (`v0.8.32` on
+2026-09-28) and is not admitted. See
 [AI Gateway](./model-gateways.md).
 
 ## Desktop, extension, and additional backend generators

package/docs/model-gateways.md

@@ -23,10 +23,10 @@ These kits ship as source-ready. That is not Workspai release admission and
 does not label them `stable` or `qualified`. SDK `1.x` stability is an upstream
 classification, not a Workspai release gate.
 
-| Kit | Runtime | Official Client SDK | Reviewed |
-| --- | --- | --- | --- |
-| `gateway.openrouter.typescript` | Node.js `>=20` | `@openrouter/sdk` `1.3.11` | 2026-09-21 |
-| `gateway.openrouter.python` | Python `>=3.10` | `openrouter` `1.2.11` | 2026-09-21 |
+| Kit                             | Runtime         | Official Client SDK        | Reviewed   |
+| ------------------------------- | --------------- | -------------------------- | ---------- |
+| `gateway.openrouter.typescript` | Node.js `>=20`  | `@openrouter/sdk` `1.3.33` | 2026-09-28 |
+| `gateway.openrouter.python`     | Python `>=3.10` | `openrouter` `1.2.32`      | 2026-09-28 |
 
 Aliases: `openrouter.typescript`, `gateway.openrouter.ts`, `openrouter-typescript`,
 and the matching Python aliases. Bare `openrouter` is not an alias.
@@ -78,9 +78,9 @@ same commit before a Workspai release can present these kits as qualified.
 ## Intentionally unsupported languages
 
 The official Go SDK (`github.com/OpenRouterTeam/go-sdk`) was evaluated on
-2026-09-21. The upstream README classifies the current `0.x` line as beta and
+2026-09-28. The upstream README classifies the current `0.x` line as beta and
 warns that breaking changes may ship in minor releases. proxy.golang.org
-reported `v0.8.11`. Go is excluded from the stable Create surface until
+reported `v0.8.32`. Go is excluded from the stable Create surface until
 upstream classifies a release line as stable.
 
 The OpenRouter Agent SDK (`@openrouter/agent`) belongs to the agent-runtime
@@ -106,22 +106,22 @@ Routing and privacy policy live in source-controlled `gateway.policy.json`.
 Field names match the pinned SDK for that language. Unknown fields are rejected
 fail-closed in both languages. Supported controls:
 
-| Official field | TypeScript policy | Python policy | Validation |
-| --- | --- | --- | --- |
-| model fallbacks | `models` | `models` | unique non-empty strings; must not repeat `OPENROUTER_MODEL` |
-| `allowFallbacks` | `allowFallbacks` | `allow_fallbacks` | boolean when present |
-| `order` / `only` / `ignore` | same camelCase | same snake_case | unique non-empty strings; `only`∩`ignore` empty; `order` must not include ignored providers |
-| `sort` | string or `{ by, partition }` | string or `{ by, partition }` | `price`/`throughput`/`latency`/`exacto`; `partition` is `model` or `none` |
-| `quantizations` | camelCase list | snake_case list | pinned SDK enum values |
-| `requireParameters` | `requireParameters` | `require_parameters` | boolean when present |
-| `zdr` | `zdr` | `zdr` | boolean; cannot combine with `dataCollection`/`data_collection` `allow` |
-| `dataCollection` | `dataCollection` | `data_collection` | `allow` or `deny` |
-| `enforceDistillableText` | `enforceDistillableText` | `enforce_distillable_text` | boolean when present; both pinned SDKs expose this field |
-| `maxPrice` | `maxPrice` | `max_price` | object with `prompt`/`completion`/`request`/`image`/`audio` as non-negative numeric strings that remain finite after numeric interpretation |
-| `preferredMinThroughput` | number or `{ p50, p75, p90, p99 }` | same | finite non-negative numbers |
-| `preferredMaxLatency` | number or `{ p50, p75, p90, p99 }` | same | finite non-negative numbers |
-| timeout | `timeoutMs` | `timeout_ms` | integer 1–600000; env overrides with a digit string |
-| attribution | `httpReferer`, `appTitle` | `http_referer`, `x_open_router_title` | strings when present |
+| Official field              | TypeScript policy                  | Python policy                         | Validation                                                                                                                                  |
+| --------------------------- | ---------------------------------- | ------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------- |
+| model fallbacks             | `models`                           | `models`                              | unique non-empty strings; must not repeat `OPENROUTER_MODEL`                                                                                |
+| `allowFallbacks`            | `allowFallbacks`                   | `allow_fallbacks`                     | boolean when present                                                                                                                        |
+| `order` / `only` / `ignore` | same camelCase                     | same snake_case                       | unique non-empty strings; `only`∩`ignore` empty; `order` must not include ignored providers                                                 |
+| `sort`                      | string or `{ by, partition }`      | string or `{ by, partition }`         | `price`/`throughput`/`latency`/`exacto`; `partition` is `model` or `none`                                                                   |
+| `quantizations`             | camelCase list                     | snake_case list                       | pinned SDK enum values                                                                                                                      |
+| `requireParameters`         | `requireParameters`                | `require_parameters`                  | boolean when present                                                                                                                        |
+| `zdr`                       | `zdr`                              | `zdr`                                 | boolean; cannot combine with `dataCollection`/`data_collection` `allow`                                                                     |
+| `dataCollection`            | `dataCollection`                   | `data_collection`                     | `allow` or `deny`                                                                                                                           |
+| `enforceDistillableText`    | `enforceDistillableText`           | `enforce_distillable_text`            | boolean when present; both pinned SDKs expose this field                                                                                    |
+| `maxPrice`                  | `maxPrice`                         | `max_price`                           | object with `prompt`/`completion`/`request`/`image`/`audio` as non-negative numeric strings that remain finite after numeric interpretation |
+| `preferredMinThroughput`    | number or `{ p50, p75, p90, p99 }` | same                                  | finite non-negative numbers                                                                                                                 |
+| `preferredMaxLatency`       | number or `{ p50, p75, p90, p99 }` | same                                  | finite non-negative numbers                                                                                                                 |
+| timeout                     | `timeoutMs`                        | `timeout_ms`                          | integer 1–600000; env overrides with a digit string                                                                                         |
+| attribution                 | `httpReferer`, `appTitle`          | `http_referer`, `x_open_router_title` | strings when present                                                                                                                        |
 
 `exacto` is part of the pinned SDK `ProviderSort` enums, not a Workspai extension.
 Malformed JSON and wrong types raise `GatewayConfigurationError` before SDK
@@ -175,7 +175,7 @@ SDK `cancel()`. A cleanup failure does not replace the primary error or convert
 a successful completion into an error.
 
 The TypeScript starter forwards `AbortSignal` through `chat.send` options,
-including an already-aborted signal. Pinned `@openrouter/sdk` `1.3.11`
+including an already-aborted signal. Pinned `@openrouter/sdk` `1.3.33`
 `EventStream` extends `ReadableStream` and does not expose a dedicated `close()`.
 Cancellation is `ReadableStream` async-iterator `return()` → `cancel()`. If the
 stream has already errored, `cancel()` may be a no-op; the gateway still drops

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "workspai",
-  "version": "0.78.0",
+  "version": "0.79.1",
   "type": "module",
   "description": "Open-source workspace intelligence CLI for software systems: create, adopt, govern, verify, and align polyglot workspaces for humans, CI, IDEs, and AI agents.",
   "keywords": [
@@ -216,7 +216,7 @@
     "@vitest/ui": "^4.1.11",
     "eslint": "^9.39.2",
     "husky": "^9.1.7",
-    "lint-staged": "^15.5.2",
+    "lint-staged": "^16.4.0",
     "prettier": "^3.6.2",
     "size-limit": "12.1.0",
     "tsup": "^8.5.1",

package/dist/analyze-X76YV667.js

@@ -1 +0,0 @@
-export{b as printAnalyzeReport,a as runAnalyze}from'./chunk-QMNJTI65.js';

package/dist/autopilot-release-6OV2FLRS.js

@@ -1 +0,0 @@
-export{b as AUTOPILOT_RELEASE_ALIAS_FILENAME,a as AUTOPILOT_RELEASE_LAST_RUN_FILENAME,c as runAutopilotRelease}from'./chunk-YHQK4S3H.js';