ur-agent 1.45.3 → 1.45.5

package/docs/AGENT_FEATURES.md

@@ -9,6 +9,15 @@ reproducible autonomous software engineering agent: every substantial task can
 be driven as `spec -> plan -> patch -> test -> report -> rollback`, with the
 spec as the durable source of truth and command evidence as the success gate.
 
+## v1.45.5 Additions
+
+- Ollama Cloud requests have bounded response-header and streaming phases and
+  do not amplify a deliberate stream timeout through fallback retries. Local
+  model timing and explicit timeout overrides are unchanged.
+- The verifier now recognizes terminal promises such as "Let me create it
+  now" or "I will run the tests now" and requires the corresponding successful
+  mutation or Bash call before the turn may complete.
+
 ## Commands
 
 ```sh
@@ -81,6 +90,24 @@ The optional `provider.fallback` value is used only to print an explicit
 recovery recommendation in provider diagnostics. It never silently or
 automatically changes the active provider.
 
+## v1.45.4 Additions
+
+- Fresh workspaces require a provider/model choice before their first
+  interactive session and persist the validated pair locally.
+- Fresh headless workspaces fail before model execution unless a model is
+  supplied explicitly or by workspace/managed configuration.
+- Resume, initialization-only, and explicit model paths remain uninterrupted;
+  AutoApprove behavior is unchanged.
+
+## v1.45.3 Additions
+
+| Addition | Surface | What it adds |
+| --- | --- | --- |
+| Deterministic slash registry | `src/commands.ts`, `test/commandRegistryIntegrity.test.ts` | Resolves bundled, plugin, project, workflow, and built-in commands by explicit source priority; rejects duplicate canonical tokens, removes only conflicting aliases, validates every lazy loader, and keeps the technical command catalog complete. |
+| Unified sandbox command | `/sandbox [status\|check\|init\|eval\|exclude]` | One interactive slash command now owns both the settings UI and text subcommands; the shell-facing `ur sandbox` implementation is shared rather than separately registered. |
+| Actionable CI cwd handling | `ur ci-loop --cwd <path>` | Reports the absolute execution directory, retains assertion and stack context, and stops "No tests found" after one attempt without wasting a fix-agent run. |
+| Explicit worktree completion | `/debug-v2`, `/refactor`, `/paper-implementation`, `/benchmark`, `/security-review`, `/dockerize`, `/latex-paper`, `/batch` | Keeps changes local, asks before the final full suite, and never commits, pushes, or opens a PR unless separately requested. |
+
 ## v1.25.x Additions
 
 | Addition | Surface | What it adds |
@@ -286,7 +313,7 @@ ur escalate plan "debug the race condition in the scheduler"
 ur escalate run "refactor the cache layer" --force-oracle
 ur escalate oracle "is this lock-free queue correct?"
 ur arena "implement the rate limiter" --agents 3 --apply
-ur ci-loop --command "bun test" --max-attempts 3
+ur ci-loop --command "bun test" --cwd . --max-attempts 3
 ur artifacts capture-diff
 ur artifacts capture-tests --command "bun test"
 ur artifacts approve 1

package/docs/AGENT_TRENDS.md

@@ -48,7 +48,7 @@ ur context-pack remember --decision "Use manifest commands first"
 ur context-pack compress
 ur acp serve --port 8123
 ur exec "add tests for the parser" --concurrency 4 --json
-ur ci-loop --command "bun test" --dry-run
+ur ci-loop --command "bun test" --cwd . --dry-run
 ur artifacts capture-diff
 ur artifacts capture-tests --command "bun test"
 ur claim-ledger validate
@@ -87,7 +87,7 @@ Inside an interactive session:
 | Spec-driven development | Covered | `ur spec` scaffolds requirements/design/tasks under `.ur/specs/`, tracks phase/approvals, and runs the Spec Kit / Kiro task list one task at a time | Add bidirectional sync with an external `specs/` directory |
 | Capability-aware model escalation | Covered | `ur escalate` selects fast/oracle tiers from `model-doctor`, runs routine work fast, and auto-escalates hard/failed work to the strong local model | Learn per-model success rates to tune the difficulty threshold |
 | Best-of-N agent judging | Covered | `ur arena` runs N agents per task in isolated worktrees and judges diffs with the self-review gate; winner is selectable/appliable | Add an optional model judge alongside the deterministic scorer |
-| Self-healing CI | Covered | `ur ci-loop` runs a command, summarizes failures, invokes a fix agent, and re-runs with bounded retries; commits/pushes are self-review gated | Wire to `ur trigger` so a failed CI webhook auto-launches the loop |
+| Self-healing CI | Covered | `ur ci-loop` reports its resolved cwd, preserves assertion/stack context, stops no-test configuration failures before invoking a fixer, and re-runs real failures with bounded retries; commits/pushes require explicit flags and are self-review gated | Wire to `ur trigger` so a failed CI webhook can explicitly launch the loop |
 | Verifiable artifacts | Covered | `ur artifacts` records plans/diffs/test-runs with approve/reject/feedback under `.ur/artifacts/`; comments steer active background agents through stream-json inbox injection | Attach browser-QA screenshots and link artifacts to claim-ledger entries |
 | Native IDE review | Covered | `ur ide diff` writes `.ur/ide/diffs/` bundles and `extensions/vscode-ur-inline-diffs/` provides a VS Code tree/webview/comment surface | Add JetBrains packaging if demand appears |
 | ACP / IDE agent server | Covered | `ur acp serve|stop|status` exposes an HTTP+JSON-RPC agent server so VS Code/Cursor/Zed extensions can list tools, call tools, and submit tasks | Ship reference editor extensions |
@@ -105,7 +105,7 @@ ur spec run demo --all --dry-run
 ur arena "implement a debounce helper" --agents 2 --dry-run
 ur escalate run "refactor the cache layer" --force-oracle --dry-run
 ur test-first --dry-run
-ur ci-loop --command "bun test" --dry-run
+ur ci-loop --command "bun test" --cwd . --dry-run
 ur artifacts capture-tests --command "bun test"
 ```
 

package/docs/CONFIGURATION.md

@@ -77,6 +77,15 @@ incompatible saved model instead of silently carrying it across providers. The
 saved provider/model pair controls the runtime backend for the next agent
 request; Ollama is only used when `ollama` is the selected provider.
 
+The same provider-first picker is mandatory on the first interactive run in a
+workspace with no model in `.ur/settings.json` or `.ur/settings.local.json`.
+The result is validated and written to the gitignored local settings file.
+User-global and built-in model defaults do not silently select a model for a
+new folder. `--model`, `OLLAMA_MODEL`, `UR_MODEL`, agent configuration,
+`--settings`, managed settings, and resumed sessions are deliberate selections
+and therefore do not open the startup picker. Fresh `-p` runs without one of
+those inputs fail with an actionable message before making a model request.
+
 Use this to inspect the active runtime path:
 
 ```sh
@@ -127,8 +136,8 @@ UR_MODEL=qwen2.5-coder:7b
 ```
 
 `OLLAMA_MODEL` selects the model name and takes precedence over `UR_MODEL` only
-for Ollama runtime sessions. If neither is set, UR lets its Ollama router choose
-from the model list advertised by the configured Ollama app.
+for Ollama runtime sessions. If neither is set and the workspace has no saved
+model, interactive startup asks and headless startup requires `--model`.
 
 ### Discovering LAN Ollama servers
 

package/docs/DEVELOPMENT.md

@@ -56,6 +56,11 @@ context.
 
 The GitHub install path uses the bundled launcher in `dist/cli.js`, so `bun run bundle` must be run before packaging or pushing a release. `bun run release:check` verifies that `package.json`, `bunfig.toml`, the bundle, docs, and `node ./bin/ur.js --version` agree. The GitHub workflow keeps production bundle, release, package, and global-install checks behind the Bun test step; do not publish, tag, or push release artifacts until that workflow is green.
 
+For a patch release, keep the version synchronized in `package.json`,
+`bunfig.toml`, the VS Code extension manifest, the JetBrains Gradle build,
+`documentation/index.html`, `technical/README.md`, and `CHANGELOG.md`. Rebuild
+`dist/cli.js` after changing the version.
+
 ## Build
 
 ```sh
@@ -73,6 +78,7 @@ only the root README. Check:
 - `CHANGELOG.md`
 - `docs/`
 - `documentation/`
+- `technical/`
 - `examples/`
 - extension and plugin README files when the feature affects them
 
@@ -80,6 +86,11 @@ For top-level commands, also update the static documentation site command data
 in `documentation/app.js` and any relevant tutorial section in
 `documentation/index.html`.
 
+For slash-command changes, update `technical/03-slash-commands.md` and run
+`bun test test/commandRegistryIntegrity.test.ts`. That test verifies unique
+canonical names and aliases, non-empty descriptions, loadable implementations,
+and coverage of every visible shipped command in the technical reference.
+
 For command surfaces that write `.ur/` state, update configuration and
 validation docs with the generated files and cleanup expectations.
 

package/docs/TROUBLESHOOTING.md

@@ -182,12 +182,13 @@ ur sandbox check
   gates.
 
 ```sh
-ur ci-loop --command "bun test" --max-attempts 3
+ur ci-loop --command "bun test" --cwd . --max-attempts 3
 ur test-first detect
 ```
 
 If the runner says "No tests found", check the working directory printed by
-the result and run from the test root or pass it explicitly:
+the result and run from the test root or pass it explicitly. This configuration
+failure stops after the first attempt and does not invoke a fix agent:
 
 ```sh
 ur ci-loop --command "bun test" --cwd ./packages/app --max-attempts 3

package/docs/USAGE.md

@@ -8,6 +8,13 @@ UR is a terminal agent. Running `ur` opens an interactive session in the current
 ur
 ```
 
+On the first interactive run in a folder without a workspace model, UR opens
+the provider-first model picker before the REPL. A validated choice is saved to
+`.ur/settings.local.json` and reused in that workspace. A global user model is
+not silently copied into new folders. Shared `.ur/settings.json`, managed,
+`--settings`, agent, CLI, environment, and resumed-session choices remain
+explicit inputs and skip this one-time picker.
+
 Use interactive mode for iterative coding, debugging, research, and repository exploration. The session can read project instructions, use tools, call slash commands, and keep resumable conversation history.
 
 Useful options:
@@ -29,9 +36,13 @@ UR also accepts custom "Other" answers when the dialog includes one.
 Print mode is useful for scripts and shell pipelines:
 
 ```sh
-ur -p "write a changelog entry for the current diff"
+ur -p --model qwen2.5-coder:7b "write a changelog entry for the current diff"
 ```
 
+In a fresh workspace without a configured model, print mode exits before any
+model request and tells the caller to pass `--model <model>` or run interactive
+setup. This keeps automation deterministic instead of selecting a default.
+
 Output formats:
 
 ```sh
@@ -57,8 +68,9 @@ in this order:
 1. `OLLAMA_MODEL`
 2. `UR_MODEL`
 
-If neither variable is set and `ollama` is the selected provider, UR lets its
-Ollama router choose from the models exposed by your local Ollama app.
+If neither variable is set, a fresh workspace requires an interactive choice.
+After the provider/model pair has been stored locally, that exact pair is used
+for subsequent sessions.
 
 You can also choose the model for a single session:
 
@@ -88,6 +100,12 @@ use `ollama.host` in settings if you want plain `ur` to default to a LAN host.
 Models exposed by the chosen Ollama app are valid, including local models and
 Ollama Cloud-backed models.
 
+Ollama Cloud models use a 120-second default bound for both response-header
+waiting and stream consumption. A deliberate stream deadline is returned
+directly instead of being replayed through the non-streaming fallback. Local
+Ollama models keep the five-minute default. Set `API_TIMEOUT_MS` to explicitly
+override either default for the current process.
+
 UR-Nexus also has explicit provider commands for legal access paths:
 
 ```sh
@@ -241,7 +259,9 @@ UR includes slash commands and CLI subcommands for common workflows:
 - `ur escalate ...` to plan, run, or ask an oracle model for hard tasks
 - `ur arena ...` to run multiple agents on the same task and select a winner
 - `ur test-first ...` to detect compile/test/lint commands, store failure traces, and install after-edit gates
-- `ur ci-loop ...` to run tests, repair failures, and rerun with a bounded loop
+- `ur ci-loop ...` to run tests in an explicit working directory, repair real
+  failures, and rerun with a bounded loop. A "No tests found" result stops
+  after one attempt and reports how to correct `--cwd`.
 - `ur artifacts ...` to capture reviewable diffs, test runs, notes, and feedback
 - `ur ide diff ...` to capture editor-readable inline diff bundles
 - `ur acp ...` to start/stop/status the Agent Communication Protocol server for IDE extensions
@@ -334,7 +354,7 @@ ur context-pack remember --accepted "Use p-map for concurrency" --scope project
 ur context-pack compress
 ur acp serve --port 8123
 ur exec "add tests for the parser" --concurrency 4 --json
-ur ci-loop --command "bun test" --dry-run
+ur ci-loop --command "bun test" --cwd . --dry-run
 ur artifacts capture-diff
 ur bg run "fix the flaky parser test" --worktree --dry-run
 ur worktree list

package/docs/VALIDATION.md

@@ -19,10 +19,42 @@ You need:
 
 ```sh
 ur --version
-# expected: the version from package.json, e.g. "1.35.1 (UR-Nexus)"
+# expected for this release: "1.45.5 (UR-Nexus)"
 ```
 
-## 0.1 Permission safety and context pack (1.19.0)
+## 0.1 First-workspace model selection (1.45.4)
+
+In a new temporary directory, run `ur`. Expected: before the REPL appears, UR
+opens the provider-first picker. After selecting a connected provider and one
+of its models, `.ur/settings.local.json` contains both `provider.active` and
+the selected model. A second `ur` run in the same directory reuses that local
+pair without showing the startup picker.
+
+Before making the interactive choice, the equivalent headless command must
+exit non-zero without contacting a model:
+
+```sh
+ur -p "say hi"
+```
+
+Expected: `No model has been selected for this workspace`. Supplying
+`--model <model>`, `OLLAMA_MODEL`, or `UR_MODEL` bypasses the gate for that run.
+
+### 0.1.1 Ollama Cloud latency containment (1.45.5)
+
+Run the deterministic regression coverage:
+
+```sh
+bun test test/ollamaTimeout.test.ts
+```
+
+Expected: the cloud-model timeout, override-precedence, stream-deadline, and
+fallback-suppression cases pass. A model ending in `:cloud` defaults to 120
+seconds for response headers and streaming; a local model retains 300 seconds.
+`API_TIMEOUT_MS` wins over both. When a cloud stream reaches its deliberate
+deadline, the request fails once instead of starting a non-streaming replay.
+
+## 0.2 Permission safety and context pack (1.19.0)
 
 In a project checkout:
 
@@ -47,7 +79,7 @@ Expected:
 - `remember` appends task memory entries under `.ur/context/task-memory.jsonl`.
 - `compress` writes `.ur/context/compressed.md`.
 
-## 0.2 Test-first execution loop (1.18.0)
+## 0.3 Test-first execution loop (1.18.0)
 
 In a project checkout:
 
@@ -74,7 +106,7 @@ ur test-first --max-attempts 1
 Expected: a non-zero command creates a log under
 `.ur/test-first/traces/`, and the command reports `exhausted`, not `passed`.
 
-## 0.3 Reliable repo editing (1.17.0)
+## 0.4 Reliable repo editing (1.17.0)
 
 In a disposable checkout:
 
@@ -92,7 +124,7 @@ Expected:
 - If the check command exits non-zero, every touched file is restored and the
   JSON result reports `"rolledBack": true`.
 
-## 0.4 Network Ollama discovery (1.16.0)
+## 0.5 Network Ollama discovery (1.16.0)
 
 With at least one other Ollama server reachable on your LAN:
 
@@ -168,6 +200,11 @@ Expected:
   UR_VERIFIER_MODE=strict ur  # default, false claim rejected
   ```
 
+Also prompt the model to end with `Let me create index.html now.` while
+forbidding tool calls. Strict mode must inject a correction and continue rather
+than accepting the promise as completion. Conditional text such as `If you
+approve, I'll create it` must not trigger this gate.
+
 ## 3. L1 loop detector fires
 
 ```text
@@ -292,6 +329,9 @@ ur artifacts capture-tests --command "bun test"
 ```
 
 Expected: no `unknown option` or `too many arguments` parser errors.
+The CI-loop result should include the resolved working directory. If that
+directory has no matching tests, the run should stop after its first attempt
+with `--cwd` guidance and must not launch a fix agent.
 
 ## What to do if any step fails
 

package/docs/providers.md

@@ -431,6 +431,13 @@ Local/server providers use their normal endpoints:
 - llama.cpp server mode: `http://localhost:8080/v1`
 - vLLM server mode: `http://localhost:8000/v1`
 
+Ollama models whose names end in `:cloud` use a 120-second default for the
+response-header phase and a 120-second total stream deadline. UR does not
+automatically replay a cloud request after that stream deadline, preventing a
+bounded failure from expanding into the shared non-streaming fallback and
+retry chain. Local Ollama models retain the five-minute default. A positive
+`API_TIMEOUT_MS` or explicit request timeout overrides these defaults.
+
 ## Optional Live Provider Smoke
 
 `bun run provider:smoke` runs optional live checks only for providers with the

package/documentation/app.js

@@ -2,7 +2,7 @@ const featureGroups = [
   {
     title: 'Core agent runtime',
     tags: ['interactive', 'headless', 'models'],
-    text: 'Interactive terminal sessions, one-shot print mode, JSON and stream-json output, resumable conversations, custom agents, model routing, local runtimes, and legal provider adapters.',
+    text: 'Interactive terminal sessions with first-workspace provider/model selection, deterministic one-shot print mode, JSON and stream-json output, resumable conversations, custom agents, model routing, local runtimes, and legal provider adapters.',
     commands: ['ur', 'ur -p', 'ur --resume', 'ur --continue', 'ur --model <model>', 'ur provider'],
   },
   {
@@ -72,15 +72,15 @@ const commands = [
     name: 'ur',
     category: 'Core',
     aliases: [],
-    summary: 'Start an interactive UR-Nexus session in the current workspace.',
+    summary: 'Start an interactive session; a fresh workspace must choose and locally persist a validated provider/model pair first.',
     examples: ['ur', 'ur --model qwen3-coder:480b-cloud', 'ur --continue', 'ur --resume'],
   },
   {
     name: 'ur -p',
     category: 'Core',
     aliases: ['--print'],
-    summary: 'Run one prompt headlessly and exit. This is the base mode for scripts, CI, evals, triggers, SDK calls, and A2A tasks.',
-    examples: ['ur -p "Summarize this repository"', 'ur -p --output-format json "Review the current diff"', 'ur -p --json-schema \'{"type":"object"}\' "Return structured output"'],
+    summary: 'Run one prompt headlessly and exit. Fresh workspaces require an explicit or workspace-configured model instead of loading a default.',
+    examples: ['ur -p --model qwen3-coder:480b-cloud "Summarize this repository"', 'ur -p --model qwen3-coder:480b-cloud --output-format json "Review the current diff"', 'ur -p --model qwen3-coder:480b-cloud --json-schema \'{"type":"object"}\' "Return structured output"'],
   },
   {
     name: 'acp',
@@ -191,8 +191,8 @@ const commands = [
     name: 'ci-loop',
     category: 'Automation',
     aliases: ['heal'],
-    summary: 'Run a build or test command, summarize failures, invoke a fix agent, and rerun with a bounded retry budget.',
-    examples: ['ur ci-loop --command "bun test" --dry-run', 'ur ci-loop --command "bun test" --max-attempts 3', 'ur ci-loop --from-log failure.log --dry-run', 'ur ci-loop --command "bun test" --commit'],
+    summary: 'Run a build or test command in an explicit working directory, preserve useful failure context, and rerun real failures with a bounded fix budget. No-test configuration failures stop immediately.',
+    examples: ['ur ci-loop --command "bun test" --cwd . --dry-run', 'ur ci-loop --command "bun test" --cwd ./packages/app --max-attempts 3', 'ur ci-loop --from-log failure.log --cwd . --dry-run', 'ur ci-loop --command "bun test" --cwd . --json'],
   },
   {
     name: 'context-pack',
@@ -264,13 +264,6 @@ const commands = [
     summary: 'Track persistent long-horizon objectives with notes, linked workflows, and resumable execution.',
     examples: ['ur goal add release-docs --objective "Ship professional docs" --workflow docs', 'ur goal note release-docs --note "Command reference drafted"', 'ur goal resume release-docs --dry-run', 'ur goal done release-docs'],
   },
-  {
-    name: 'update',
-    category: 'Ops',
-    aliases: ['upgrade'],
-    summary: 'Check npm for UR-Nexus updates.',
-    examples: ['ur update', 'ur upgrade'],
-  },
   {
     name: 'connect',
     category: 'Providers',
@@ -432,6 +425,41 @@ const commands = [
     summary: 'Parse GitHub, Slack, or generic webhook payloads and optionally launch a headless UR run.',
     examples: ['ur trigger parse --file payload.json --source github', 'ur trigger run --file payload.json --keyword /ur --dry-run', 'ur trigger run --file slack.json --max-turns 8'],
   },
+  {
+    name: 'audit',
+    category: 'Verification',
+    aliases: [],
+    summary: 'Export the hash-chained project audit trail or verify it for tampering.',
+    examples: ['ur audit export --format jsonl', 'ur audit export --format csv --out audit.csv', 'ur audit verify'],
+  },
+  {
+    name: 'cloud',
+    category: 'Agent Platform',
+    aliases: [],
+    summary: 'Run detached best-of-N tasks, inspect their results, and explicitly apply a selected winner.',
+    examples: ['ur cloud run "speed up parser" --attempts 3', 'ur cloud list', 'ur cloud show <id>', 'ur cloud apply <id>'],
+  },
+  {
+    name: 'recipe',
+    category: 'Automation',
+    aliases: ['recipes'],
+    summary: 'Create and run structured-output playbooks whose child result must satisfy the recipe schema.',
+    examples: ['ur recipe init triage', 'ur recipe list', 'ur recipe run triage "login returns 500"'],
+  },
+  {
+    name: 'thread',
+    category: 'Sessions',
+    aliases: ['threads'],
+    summary: 'Share a session transcript through the local artifacts server or list shared threads.',
+    examples: ['ur thread share', 'ur thread list', 'ur thread share <session-id>'],
+  },
+  {
+    name: 'wiki',
+    category: 'Knowledge',
+    aliases: ['repo-wiki'],
+    summary: 'Generate a living repository wiki and map, inspect status, or install its maintenance hook.',
+    examples: ['ur wiki generate', 'ur wiki map', 'ur wiki status', 'ur wiki install-hook'],
+  },
   {
     name: 'update',
     category: 'Ops',
@@ -472,7 +500,7 @@ const slashGroups = [
   {
     title: 'Agent skills',
     items: ['/debug-v2', '/refactor', '/paper-implementation', '/benchmark', '/security-review', '/dockerize', '/latex-paper', '/simplify', '/debug'],
-    text: 'Bundled slash skills that dispatch focused agent work in isolated git worktrees with clean commits and PR output.',
+    text: 'Bundled slash skills that dispatch focused work in isolated git worktrees, keep changes local, ask before the final full suite, and publish only after a separate explicit request.',
   },
   {
     title: 'Memory and evidence',
@@ -486,8 +514,8 @@ const slashGroups = [
   },
   {
     title: 'Models, tools, and interop',
-    items: ['/model', '/model-doctor', '/model-route', '/escalate', '/mcp', '/plugin', '/skills', '/sdk', '/a2a-card'],
-    text: 'Pick models, inspect capabilities, escalate to oracle models, manage MCP/plugin extensions, and expose interop surfaces.',
+    items: ['/model', '/model-doctor', '/model-route', '/escalate', '/mcp', '/plugin', '/skills', '/skill', '/sdk', '/a2a-card'],
+    text: 'Pick models, inspect capabilities, manage MCP/plugin extensions, browse prompt skills with /skills, run executable workflows with /skill, and expose interop surfaces.',
   },
   {
     title: 'Security operations',
@@ -502,7 +530,7 @@ const slashGroups = [
   {
     title: 'Preferences and environment',
     items: ['/config', '/permissions', '/sandbox', '/mode', '/theme', '/vim', '/terminal-setup', '/usage', '/ide'],
-    text: 'Configure permissions, sandboxing, UI preferences, IDE integration, and runtime environment details.',
+    text: 'Configure permissions and UI preferences; /sandbox combines its interactive settings with status, check, init, eval, and exclude operations.',
   },
 ];
 
@@ -662,8 +690,8 @@ const examples = [
   },
   {
     title: 'Self-healing CI',
-    text: 'Run a test command, summarize failures, attempt a bounded fix loop, and rerun.',
-    code: 'ur ci-loop --command "bun test" --dry-run\nur ci-loop --command "bun test" --max-attempts 3',
+    text: 'Run in an explicit cwd, preserve useful assertion and stack context, and attempt a bounded fix loop only for actionable failures.',
+    code: 'ur ci-loop --command "bun test" --cwd . --dry-run\nur ci-loop --command "bun test" --cwd ./packages/app --max-attempts 3',
   },
   {
     title: 'Test-first command evidence',

package/documentation/index.html

@@ -44,7 +44,7 @@
     <main id="content" class="content">
       <header class="topbar">
         <div>
-          <p class="eyebrow">Version 1.45.3</p>
+          <p class="eyebrow">Version 1.45.5</p>
           <h1>UR-Nexus Documentation</h1>
           <p class="lead">A practical, tutorial-style reference for installing, configuring, automating, extending, and operating UR-Nexus.</p>
         </div>
@@ -89,6 +89,10 @@
             <strong>Test-first execution</strong>
             <span><code>ur test-first</code> detects compile/test/lint commands, stores failure traces, and installs edit-time verifier gates.</span>
           </article>
+          <article>
+            <strong>Deterministic commands</strong>
+            <span>The external runtime exposes 150 commands and 217 unique slash tokens; registry tests reject ambiguous names, broken loaders, and undocumented visible commands.</span>
+          </article>
           <article>
             <strong>Safety and context</strong>
             <span><code>ur safety</code> evaluates command risk. <code>ur context-pack</code> writes architecture context and task memory.</span>
@@ -122,13 +126,13 @@ ur --model qwen3-coder:480b-cloud</code></pre>
           <article class="step">
             <span>3</span>
             <h3>Run once in scripts</h3>
-            <pre><code>ur -p "Summarize this repository"
-ur -p --output-format json "Review the current diff"</code></pre>
+            <pre><code>ur -p --model qwen3-coder:480b-cloud "Summarize this repository"
+ur -p --model qwen3-coder:480b-cloud --output-format json "Review the current diff"</code></pre>
           </article>
         </div>
         <div class="callout">
           <strong>Runtime requirement</strong>
-          <p>UR expects Bun. For local-first work, run a local Ollama app/server at <code>http://localhost:11434/api</code>. Set <code>OLLAMA_MODEL</code> or <code>UR_MODEL</code> to force a model route.</p>
+          <p>UR expects Bun. On the first interactive run in a workspace, choose a provider and model; UR validates and saves the pair locally before starting the session. Fresh print-mode workspaces require <code>--model</code> or another explicit model source.</p>
         </div>
       </section>
 
@@ -159,9 +163,14 @@ ur config set provider.fallback ollama</code></pre>
           </article>
           <article>
             <h3>Status bar and updates</h3>
-            <pre><code>Ollama | llama3 | ask | main | update 1.37.3 available</code></pre>
+            <pre><code>Ollama | llama3 | ask | main | update 1.45.5 available</code></pre>
             <p>The interactive status bar shows only important runtime state: provider, model, mode, branch, active tasks, checks status when known, and update availability. It is hidden in CI, dumb terminals, and print mode.</p>
           </article>
+          <article>
+            <h3>Ollama Cloud latency bounds</h3>
+            <pre><code>API_TIMEOUT_MS=120000 ur</code></pre>
+            <p>Cloud-tagged Ollama models default to bounded response-header and streaming phases. A stream deadline is surfaced directly instead of starting a non-streaming retry chain. Local models retain their longer default, and <code>API_TIMEOUT_MS</code> remains an explicit override.</p>
+          </article>
         </div>
       </section>
 
@@ -396,13 +405,14 @@ ur arena "fix the parser" --agents 3 --apply</code></pre>
           <article>
             <h3>Repair failing CI in a bounded loop</h3>
             <ol>
-              <li>Run the build or test command.</li>
-              <li>Summarize the failure and launch a fix attempt.</li>
+              <li>Run the build or test command in the printed working directory.</li>
+              <li>Preserve the assertion and stack context, then launch a fix attempt only for an actionable failure.</li>
               <li>Rerun until the command passes or the retry budget is exhausted.</li>
             </ol>
-            <pre><code>ur ci-loop --command "bun test" --dry-run
-ur ci-loop --command "bun test" --max-attempts 3
-ur ci-loop --from-log failure.log --dry-run</code></pre>
+            <p>A <code>No tests found</code> result stops after the first attempt and points to <code>--cwd</code>; it does not invoke a fix agent.</p>
+            <pre><code>ur ci-loop --command "bun test" --cwd . --dry-run
+ur ci-loop --command "bun test" --cwd ./packages/app --max-attempts 3
+ur ci-loop --from-log failure.log --cwd . --dry-run</code></pre>
           </article>
 
           <article>
@@ -453,7 +463,7 @@ ur artifacts approve 1</code></pre>
             <pre><code>OLLAMA_MODEL=qwen3-coder:480b-cloud ur
 UR_MODEL=qwen3-coder:480b-cloud ur
 ur --model qwen2.5-coder:latest</code></pre>
-            <p><code>OLLAMA_MODEL</code> wins over <code>UR_MODEL</code>. Without either, UR routes from local Ollama model inventory.</p>
+            <p><code>OLLAMA_MODEL</code> wins over <code>UR_MODEL</code>. Without an explicit or workspace model, interactive startup opens the provider/model picker and fresh print mode exits before model execution.</p>
           </article>
           <article>
             <h3>Permissions</h3>
@@ -476,7 +486,7 @@ ur context-pack compress</code></pre>
 UR_VERIFIER_MODE=loose
 UR_VERIFIER_MODE=off
 UR_VERIFIER_AUTO_SUBAGENT=1</code></pre>
-            <p>L1 gates catch false done claims and loops. <code>/verify</code> manually runs deeper verification.</p>
+            <p>L1 gates catch false done claims, immediate-action promises that end without their tool call, and loops. <code>/verify</code> manually runs deeper verification.</p>
             <p>Interactive users can set <code>"verifier": { "askBeforeGates": true }</code> in <code>.ur/settings.json</code> (or via <code>ur config set verifier.askBeforeGates true</code>) to have UR ask before running project test/typecheck/lint gates after a task, instead of running them automatically. Default is <code>false</code> (auto-run gates).</p>
           </article>
           <article>
@@ -554,6 +564,12 @@ ur agent-inspect --file session.jsonl
 ur artifacts list
 ur eval report starter
 ur claim-ledger validate</code></pre>
+          </article>
+          <article>
+            <h3>CI loop reports no tests</h3>
+            <p>Use the working directory printed in the result. The loop stops after one attempt because changing code cannot repair a test-discovery cwd error.</p>
+            <pre><code>ur ci-loop --command "bun test" --cwd . --max-attempts 3
+ur ci-loop --command "bun test" --cwd ./packages/app --max-attempts 3</code></pre>
           </article>
           <article>
             <h3>New agent-platform command is missing</h3>

package/extensions/jetbrains-ur/build.gradle.kts

@@ -5,7 +5,7 @@ plugins {
 }
 
 group = "dev.urnexus"
-version = "1.45.3"
+version = "1.45.5"
 
 repositories {
     mavenCentral()

package/extensions/vscode-ur-inline-diffs/package.json

@@ -2,7 +2,7 @@
   "name": "ur-inline-diffs",
   "displayName": "UR Inline Diffs",
   "description": "Review, apply, and reject UR inline diff bundles from .ur/ide/diffs inside VS Code.",
-  "version": "1.45.3",
+  "version": "1.45.5",
   "publisher": "ur-nexus",
   "engines": {
     "vscode": "^1.92.0"

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "ur-agent",
-  "version": "1.45.3",
+  "version": "1.45.5",
   "description": "UR-Nexus — autonomous engineering workflow engine (plan, execute, test, verify, document, benchmark, reproduce)",
   "type": "module",
   "packageManager": "bun@1.3.14",