callman-mcp 0.3.0 → 0.5.0

package/dist-cjs/domain/uiTest/summarize.js

@@ -0,0 +1,155 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.imageFromDataUrl = exports.scheduleView = exports.historyRow = exports.runListRow = exports.runReportView = exports.settingsView = exports.maskedDatasetKeys = exports.flowSummaryView = void 0;
+const callman_core_1 = require("callman-core");
+const mask_js_1 = require("../../lib/mask.js");
+const truncate_js_1 = require("../../lib/truncate.js");
+// Compact UI-test views: a flow row, a run timeline with the failing steps
+// expanded, schedule rows. Screenshots never go into structuredContent —
+// they travel as MCP image blocks (tools/uiTestRuns) or as `has…` flags.
+const time = (value) => {
+    if (value === null || value === undefined || value === "")
+        return null;
+    const date = typeof value === "number" ? new Date(value) : new Date(String(value));
+    return Number.isNaN(date.getTime()) ? String(value) : date.toISOString();
+};
+const flowSummaryView = (flow, folderPath) => ({
+    id: flow.clientId,
+    name: flow.name,
+    platform: flow.platform,
+    ...(flow.description ? { description: (0, truncate_js_1.clip)(flow.description, 300) } : {}),
+    folder: folderPath ?? null,
+    tags: flow.tags ?? [],
+    testType: flow.testType ?? null,
+    severity: flow.severity ?? null,
+    version: flow.version ?? null,
+    ...(flow.reviewStatus ? { reviewStatus: flow.reviewStatus } : {}),
+    ...(flow.activeVersionNumber ? { activeVersionNumber: flow.activeVersionNumber } : {}),
+    ...(flow.groupName ? { group: flow.groupName } : {}),
+    updatedAt: time(flow.updatedAt),
+});
+exports.flowSummaryView = flowSummaryView;
+const maskedDatasetKeys = (dataset) => Object.entries(dataset ?? {}).filter(([, value]) => (0, mask_js_1.isMasked)(value)).map(([key]) => key);
+exports.maskedDatasetKeys = maskedDatasetKeys;
+const settingsView = (flow) => ({
+    ...(flow.platform === "web" ? { browserKind: flow.browserKind ?? null, headless: flow.headless ?? null } : { dismissPopups: flow.dismissPopups ?? [] }),
+    runSettings: (0, callman_core_1.resolveRunSettings)(flow.runSettings ?? null),
+});
+exports.settingsView = settingsView;
+const MAX_STEP_ROWS = 300;
+const stepRow = (step, index) => ({
+    index: index + 1,
+    command: (0, truncate_js_1.clip)(step.command ?? "", 160),
+    status: step.status,
+    durationMs: step.durationMs ?? null,
+    ...(step.error ? { error: (0, truncate_js_1.clip)(step.error, 500) } : {}),
+    ...(step.source ? { source: step.source } : {}),
+    ...(step.healed ? { healed: (0, truncate_js_1.clip)(step.healed, 200) } : {}),
+    ...(step.retries ? { retries: step.retries } : {}),
+});
+const failedStepDetail = (step, index) => ({
+    index: index + 1,
+    command: (0, truncate_js_1.clip)(step.command ?? "", 400),
+    status: step.status,
+    error: step.error ? (0, truncate_js_1.clip)(step.error, 4000) : null,
+    ...(step.source ? { source: step.source } : {}),
+    ...(step.note ? { note: (0, truncate_js_1.clip)(step.note, 500) } : {}),
+    ...(step.data !== undefined && step.data !== null ? { data: (0, truncate_js_1.jsonPreview)(step.data, 3000) } : {}),
+    ...(step.extracted ? { extracted: (0, truncate_js_1.jsonPreview)(step.extracted, 1500) } : {}),
+    ...(step.dataTruncated ? { dataTruncated: true } : {}),
+});
+const isFailed = (status) => status === "failed" || status === "error";
+const runReportView = (report, options = {}) => {
+    const steps = report.steps ?? [];
+    const counts = steps.reduce((acc, step) => ((acc[step.status] = (acc[step.status] ?? 0) + 1), acc), {});
+    const failed = steps.map((step, index) => ({ step, index })).filter(({ step }) => isFailed(step.status));
+    const withData = options.allStepData ? steps.map((step, index) => ({ step, index })).filter(({ step }) => !isFailed(step.status) && (step.data != null || step.extracted)) : [];
+    return {
+        reportId: report.id,
+        flow: { id: report.flowClientId, name: report.flowName },
+        status: report.status,
+        ...(report.executor ? { executor: report.executor } : {}),
+        ...(report.platform ? { platform: report.platform } : {}),
+        ...(report.relay?.deviceName || report.relay?.deviceId ? { device: report.relay.deviceName ?? report.relay.deviceId } : {}),
+        triggerType: report.triggerType ?? null,
+        ...(report.scheduleName ? { schedule: report.scheduleName } : {}),
+        startedAt: time(report.startedAt),
+        finishedAt: time(report.finishedAt),
+        durationMs: report.durationMs ?? null,
+        ...(report.error ? { error: (0, truncate_js_1.clip)(report.error, 4000) } : {}),
+        stepCounts: counts,
+        steps: steps.slice(0, MAX_STEP_ROWS).map(stepRow),
+        ...(steps.length > MAX_STEP_ROWS ? { stepsTruncated: steps.length } : {}),
+        ...(failed.length > 0 ? { failedSteps: failed.slice(0, 10).map(({ step, index }) => failedStepDetail(step, index)) } : {}),
+        ...(withData.length > 0 ? { stepData: withData.slice(0, 30).map(({ step, index }) => failedStepDetail(step, index)) } : {}),
+        ...(report.datasetIssues?.length ? { datasetIssues: report.datasetIssues } : {}),
+        ...(report.warnings?.length ? { warnings: report.warnings } : {}),
+        screenshots: {
+            failure: Boolean(report.failureScreenshot ?? report.hasFailureScreenshot),
+            final: Boolean(report.finalScreenshot ?? report.hasFinalScreenshot),
+        },
+        ...(options.includeNetwork
+            ? { network: (0, truncate_js_1.jsonPreview)(report.network ?? [], 12_000), ...(report.networkTruncated ? { networkTruncated: true } : {}) }
+            : (report.network?.length ?? report.networkCount)
+                ? { networkEvents: report.network?.length ?? report.networkCount }
+                : {}),
+        ...(report.fromApprovedSnapshot ? { fromApprovedSnapshot: true, flowVersionNumber: report.flowVersionNumber ?? null } : {}),
+    };
+};
+exports.runReportView = runReportView;
+const runListRow = (report) => ({
+    reportId: report.id,
+    flow: { id: report.flowClientId, name: report.flowName },
+    status: report.status,
+    ...(report.executor ? { executor: report.executor } : {}),
+    triggerType: report.triggerType ?? null,
+    startedAt: time(report.startedAt),
+    durationMs: report.durationMs ?? null,
+    ...(report.error ? { error: (0, truncate_js_1.clip)(report.error, 300) } : {}),
+    ...(report.steps ? { failedStep: (() => {
+            const index = report.steps.findIndex((step) => isFailed(step.status));
+            return index >= 0 ? { index: index + 1, command: (0, truncate_js_1.clip)(report.steps[index].command, 160) } : null;
+        })() } : {}),
+});
+exports.runListRow = runListRow;
+const historyRow = (row) => ({
+    flow: { id: row.flowClientId, name: row.flowName },
+    status: row.status,
+    source: row.triggeredBy === "schedule" ? "schedule" : row.triggeredBy ? row.triggeredBy : "desktop",
+    platform: row.platform,
+    ...(row.deviceId ? { device: row.deviceId } : {}),
+    startedAt: time(row.startedAt),
+    durationMs: row.durationMs,
+    ...(row.error ? { error: (0, truncate_js_1.clip)(row.error, 300) } : {}),
+    ...(row.reportId ? { reportId: row.reportId } : {}),
+});
+exports.historyRow = historyRow;
+const scheduleView = (schedule) => ({
+    id: schedule.id,
+    name: schedule.name,
+    flow: { id: schedule.flowClientId, name: schedule.flowName },
+    enabled: schedule.enabled,
+    triggerType: schedule.triggerType,
+    timezone: schedule.timezone,
+    ...(schedule.dailyConfig ? { dailyConfig: schedule.dailyConfig } : {}),
+    ...(schedule.weeklyConfig ? { weeklyConfig: schedule.weeklyConfig } : {}),
+    ...(schedule.monthlyConfig ? { monthlyConfig: schedule.monthlyConfig } : {}),
+    ...(schedule.customCronExpression ? { customCronExpression: schedule.customCronExpression } : {}),
+    ...(schedule.description ? { description: schedule.description } : {}),
+    ...(schedule.executionPolicy ? { executionPolicy: schedule.executionPolicy } : {}),
+    ...(schedule.failurePolicy ? { failurePolicy: schedule.failurePolicy } : {}),
+    ...(schedule.notificationConfig ? { notificationConfig: schedule.notificationConfig } : {}),
+    lastRunAt: time(schedule.lastRunAt),
+    lastRunStatus: schedule.lastRunStatus ?? null,
+    nextRunAt: time(schedule.nextRunAt),
+    ...(schedule.isRunning ? { isRunning: true } : {}),
+});
+exports.scheduleView = scheduleView;
+// `data:image/jpeg;base64,…` → an MCP image block.
+const imageFromDataUrl = (dataUrl) => {
+    if (!dataUrl)
+        return null;
+    const match = /^data:(image\/[a-z0-9.+-]+);base64,(.+)$/i.exec(dataUrl.trim());
+    return match ? { mimeType: match[1].toLowerCase(), data: match[2] } : null;
+};
+exports.imageFromDataUrl = imageFromDataUrl;

package/dist-cjs/resources/docs/overview.js

@@ -4,8 +4,8 @@ exports.OVERVIEW_DOC = void 0;
 exports.OVERVIEW_DOC = `# Working with Callman through MCP
 
 Callman is an API client and test platform (collections of HTTP requests, environments,
-multi-step test scenarios, system-design diagrams). This server exposes it to you through
-tools; everything you do goes to the same server the Callman desktop app uses, as the user
+multi-step API test scenarios, UI tests for web / Android / iOS, system-design diagrams). This
+server exposes it to you through tools; everything you do goes to the same server the Callman desktop app uses, as the user
 whose token is configured, limited to that token's permissions and the user's workspace role.
 
 ## Start here
@@ -16,7 +16,7 @@ whose token is configured, limited to that token's permissions and the user's wo
 3. Arguments are strict: a misspelled or unknown argument is an error that lists the valid
    ones — read it and retry; nothing is silently ignored.
 4. Names work as references everywhere an id does (collection, folder, request, environment,
-   scenario, diagram, workspace). Ambiguous names come back as a candidate list — pick one.
+   scenario, UI test, diagram, workspace). Ambiguous names come back as a candidate list — pick one.
 5. Preview big writes with \`dryRun: true\` (bulk create, imports, scenarios, diagrams).
 6. Deletions need \`confirm: true\` and an exact name or id. Ask the user unless they explicitly
    requested the deletion.
@@ -59,6 +59,15 @@ Multi-step flows as a graph of typed nodes (request, condition, assertion, set-v
 db, kafka, script, wait, end …). Read \`callman://docs/scenario-nodes\` before building one;
 partial node configs are merged over defaults. Runs happen on the server (\`run_scenario\`).
 
+## UI tests
+Browser and mobile-app flows (YAML: header, \`---\`, steps) — the desktop app's UI Testing area.
+Read \`callman://docs/ui-tests\` first and \`callman://docs/ui-test-steps\` for every step. Flows are
+linted before they are saved (\`validate_ui_test\`); edit steps by number with
+\`patch_ui_test_steps\`. Web flows run on the server runner when there is one; Android, iOS (and
+web without a runner) run on the user's own Callman desktop app, which must be open with
+"Let AI tools run UI tests on this computer" on — \`list_ui_test_devices\` shows what is
+available. Runs return per-step results; screenshots come back as images on request.
+
 ## Diagrams
 System-design diagrams use a fixed component catalog — \`callman://diagram/components\`.
 Build with \`create_diagram\` (auto layout), edit with \`patch_diagram\`, share with

package/dist-cjs/resources/docs/uiTests.js

@@ -0,0 +1,103 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.UI_TESTS_DOC = void 0;
+exports.UI_TESTS_DOC = `# UI tests in Callman
+
+UI tests drive a real web browser or a mobile app step by step — the same flows QA builds in
+the desktop app's UI Testing area. Every step (\`tapOn\`, \`inputText\`, \`assertVisible\`,
+\`dbQuery\` …) with its arguments is in \`callman://docs/ui-test-steps\`.
+
+## A flow
+A flow is YAML text: an optional **header**, a \`---\` line, then a **list of steps**.
+
+\`\`\`yaml
+url: https://shop.example.com        # web: where the browser starts   (mobile: appId: com.shop.app)
+---
+- runFlow: "Login"                   # precondition: another flow, by EXACT name
+- tapOn: "Catalog"
+- inputText:
+    id: "search"
+    text: "{{product}}"              # dataset placeholder
+- assertVisible:
+    text: "{{product}}"
+    label: "Search shows the product"
+- dbQuery:                           # hybrid data step (runs between UI steps)
+    connection: orders-db            # a workspace connection, by name or id
+    query: "select status from orders where id = '{{orderId}}'"
+    expect:
+      - field: "rows[0].status"
+        op: equals
+        value: "PAID"
+\`\`\`
+
+- **Platforms**: \`web\` (Playwright), \`android\` and \`ios\` (the mobile engine in the desktop app).
+  A flow has one platform; steps the platform cannot run are lint warnings.
+- **Header**: \`url\` (web) or \`appId\` (android/ios) is what matters; \`env\`, \`tags\`, \`appActivity\`,
+  \`deviceName\`, \`onFlowStart\`/\`onFlowComplete\` are optional.
+- **Preconditions / sub-flows**: \`- runFlow: "Name"\` inlines another flow of the library by its
+  exact name (recursively). The leading runFlow lines are the flow's *preconditions*
+  (create_ui_test / update_ui_test \`preconditions: ["Login"]\`). Renaming a flow breaks callers —
+  update_ui_test lists them. Names are unique per workspace.
+- **Labels**: any mapping step may carry \`label:\` — its name in reports (patch op \`setLabel\`).
+
+## Data
+- **dataset**: key → value, used as \`{{key}}\` anywhere (header included). Values whose key looks
+  secret (password, token, apiKey, pin, otp …) come back masked (\`••••ab\`) unless the token holds
+  \`environments:reveal\` and you pass \`reveal: true\`. Editing a flow without reveal is fine:
+  stored secrets are kept as they are. Never type a masked value yourself — pass the real value
+  to change a secret, or leave the key out.
+- **datasetSources**: \`{ key: { connection, query, column? } }\` — the value comes from a database
+  query at run time (a database connection of the workspace).
+- **Extracted values**: \`extract\` / \`copyTextFrom\` / data steps save \`{{vars}}\` for later steps.
+- **Hybrid steps**: \`dbQuery\`, \`apiCall\`, \`kafkaPublish\`, \`kafkaCheck\`, \`redisCheck\`, \`runScript\`
+  run outside the UI engine, in order. Connections are the workspace's (\`list_connections\`),
+  referenced by name or id.
+
+## Settings
+Per flow: \`browserKind\` (chrome / firefox / webkit) and \`headless\` for web desktop runs;
+\`dismissPopups\` (texts auto-dismissed) for mobile; \`runSettings\` { screenshotOnSuccess,
+screenshotOnFailure, autoRetryTransient, captureNetwork } — null = the shared default.
+
+## Where runs happen
+- **Server runner** (\`target: "server"\`): web flows only, when the installation runs one.
+  Schedules always use it.
+- **Your desktop app** (\`target: "desktop"\`): the Callman desktop app on the user's computer runs
+  the job — Android devices/emulators, iOS simulators/devices, local browsers. It must be open,
+  signed in, with **Settings → AI tools → "Let AI tools run UI tests on this computer"** on.
+  Desktop runs queue one at a time. \`device\`, \`browserKind\`, \`headless\` and a one-off
+  \`dataset\` override apply here.
+- \`target: "auto"\` (default): web → server runner when available, otherwise the desktop.
+- \`list_ui_test_devices\` shows the server runner, the online desktop apps and their devices.
+  \`UI_TEST_DESKTOP_OFFLINE\` means no desktop app of the user is online.
+
+## Reports
+\`run_ui_test\` / \`get_ui_test_run\` return per-step rows (command, status, duration, error, the
+sub-flow it came from), failing steps expanded with the data they returned (DB rows, API
+responses, extracted vars), dataset issues and warnings. \`includeScreenshots: true\` adds the
+failure / final screenshots as images; \`includeNetwork: true\` the captured API calls.
+\`list_ui_test_runs\` merges server reports with the history the desktop app recorded itself.
+
+## Schedules
+Web flows only, on the server runner: daily / weekly / monthly at times (HH:mm) or a cron
+expression, with retries and notifications. One call can schedule many flows. On-prem,
+schedules run the flow's APPROVED version.
+
+## Review (on-prem)
+Editing a flow makes it a draft. \`submit_ui_test\` sends it for review; an approver
+\`review_ui_test\` (approve / reject with a note). Each approval is a version
+(\`list_ui_test_versions\`, with a diff against the current content); \`restore_ui_test_version\`
+re-activates an older one. Cloud has no review workflow — flows run as saved.
+
+## Recommended workflow
+1. \`list_ui_test_devices\` — where can it run (server runner? desktop app + devices?).
+2. \`list_ui_tests\` — reuse existing flows (Login …) as preconditions instead of copying steps.
+3. Read \`callman://docs/ui-test-steps\`, draft the steps, then \`validate_ui_test\` (content +
+   platform) — or \`create_ui_test\` with \`dryRun: true\`.
+4. \`create_ui_test\` (lint errors refuse the save; warnings are reported).
+5. \`run_ui_test\` → read the failing steps; \`get_ui_test_run includeScreenshots: true\` when a
+   picture helps.
+6. Fix with \`patch_ui_test_steps\` (insert / replace / remove / move / setLabel by step number
+   from \`get_ui_test\`) and run again.
+7. Organise with folders (\`create_ui_test_folder\`, \`move_ui_tests\`), schedule web flows
+   (\`create_ui_test_schedule\`), submit for approval on-prem.
+`;

package/dist-cjs/resources/generated/aiPacks.js

@@ -1,8 +1,8 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.AI_PACKS = exports.AI_PACKS_CORE_VERSION = void 0;
-// GENERATED by scripts/build-static-resources.mjs from callman-core@1.29.0 ai-packs — do not edit.
-exports.AI_PACKS_CORE_VERSION = "1.29.0";
+// GENERATED by scripts/build-static-resources.mjs from callman-core@1.31.0 ai-packs — do not edit.
+exports.AI_PACKS_CORE_VERSION = "1.31.0";
 exports.AI_PACKS = {
     "db-queries": "<!-- GENERATED FILE — do not edit by hand. Source: callman_core (npm run ai:packs). -->\n\n# Callman data access from scripts (`pm.db.query` / `pm.redis.command`)\n\n`pm.db.query({ connectionId, query })` — ALWAYS `await`. `query` is coerced to a string, `{{tokens}}` are resolved, then dispatched by the connection's type. `connectionId` may be omitted to use the request's configured DB connection. IMPORTANT: `pm.db.query` does NOT accept a `params`/bind array — inline values (quote strings) or use `{{env.X}}` tokens; SQL placeholders like `$1` or `?` are NOT bound.\n\n## MongoDB\n`query` MUST be a JSON STRING of a command object (use `JSON.stringify`), NOT a `db.coll.find(...)` shell string. The match key is `filter` (a top-level `query` key is rejected).\n\n```js\nconst body = pm.response.json();\nconst res = await pm.db.query({\n  connectionId: \"…\",\n  query: JSON.stringify({\n    collection: \"users\",\n    operation: \"findOne\",        // find | findOne | aggregate | countDocuments | insertOne | insertMany | updateOne | updateMany | deleteOne | deleteMany | findOneAndUpdate | findOneAndReplace | findOneAndDelete\n    filter: { email: body.data.user.email },\n  }),\n});\npm.expect(res.rowCount).to.be.above(0);   // existence check\n```\n\nPer-operation fields: `find`/`findOne` use `filter` (+ optional `projection`,`sort`,`limit`,`skip`); `aggregate` uses `pipeline` (array); `insertOne` uses `document`, `insertMany` uses `documents`; `updateOne`/`updateMany`/`findOneAndUpdate` use `filter`+`update` (Mongo update operators, e.g. `{ $set: { … } }`); `deleteOne`/`deleteMany` use `filter`; `findOneAndUpdate`/`findOneAndReplace` accept `returnDocument: \"before\"|\"after\"`. A 24-hex string in a filter is auto-coerced to ObjectId (or use `{ \"$oid\": \"…\" }`). Rows are in `res.rows`; write ops surface counts in `res.rowCount`.\n\n## PostgreSQL / MySQL / Oracle (SQL)\n`query` is plain SQL text. Because `pm.db.query` binds no params, inline the value and quote strings yourself (or use a `{{token}}`):\n\n```js\nconst email = pm.response.json().data.user.email;\nconst res = await pm.db.query({\n  connectionId: \"…\",\n  query: `SELECT id FROM users WHERE email = '${email}'`,\n});\npm.expect(res.rowCount).to.be.above(0);\n```\n\nNotes: a trailing `;` is stripped (except Oracle PL/SQL blocks, which keep it). **Oracle** returns column names UPPERCASE — read `res.rows[0].EMAIL`, not `.email`. `res.rowCount` is the row count for SELECT and the affected-row count for INSERT/UPDATE/DELETE.\n\n## SQL Server (mssql)\nSingle statement per query; a trailing `;` is stripped; `GO` batch separators are REJECTED (client-tool syntax, not T-SQL). Positional params bind as named inputs: `params[0]` → `@p1`, `params[1]` → `@p2`… (scenario db node only — `pm.db.query` has no params).\n\n```sql\nSELECT TOP 1 id, amount FROM dbo.orders WHERE customer_id = @p1 ORDER BY created_at DESC\n```\n\n## Redis (`pm.redis.command`)\n`command` is a single string. Supported commands: **GET, SET, DEL, EXISTS, TTL, HGET, HSET, HGETALL, HDEL, LPUSH, RPUSH, LRANGE, LLEN, SADD, SMEMBERS, SISMEMBER, INCR, DECR, INCRBY, EXPIRE, PERSIST, TYPE, ZADD, ZRANGE, ZSCORE, SCAN**. Quote a value that contains spaces.\n\nStructured replies land in `res.result` as real values: `HGETALL` → object, `LRANGE`/`SMEMBERS`/`ZRANGE` → array, `EXISTS`/`SISMEMBER`/`EXPIRE`/`PERSIST` → boolean; `res.resultText` is the stringified form. `SCAN pattern [count]` is Callman-managed: the cursor loop runs internally (capped at 1000 keys / 100 iterations) and returns `{ keys, count, truncated }` — no raw cursors.\n\n```js\nconst res = await pm.redis.command({ connectionId: \"…\", command: `GET session:${id}` });\npm.expect(res.success).to.be.true;\n// res.result / res.resultText hold the value; res.commandName, res.key are set.\n```\n",
     "diagram-components": "<!-- GENERATED FILE — do not edit by hand. Source: callman_core (npm run ai:packs). -->\n\n# Diagram components and edge notation\n\nA diagram node's `type` MUST be one of the keys below (`callman-core` `DIAGRAM_COMPONENT_CATALOG`). Sizes are the defaults the studio uses; `container` types hold children via `parentId`.\n\n## Edge notation\n\n| Semantic | Meaning | Style tuple |\n| --- | --- | --- |\n| `sync` | Request / response call | dashed=false, strokeWidth=1.5 |\n| `async` | Fire-and-forget, event or callback | dashed=true, dashPattern=\"6 4\", strokeWidth=1.5 |\n| `fallback` | Retry, fallback or optional path | dashed=true, dashPattern=\"1 3\", strokeWidth=1.25 |\n| `bus` | Message bus or stream | dashed=false, strokeWidth=3 |\n\n## Generic (`generic`)\n\n| Type | Label | Default size | Aliases |\n| --- | --- | --- | --- |\n| `line` | Line | 160×0 | connector, edge, link |\n| `arrow` | Arrow | 160×0 | connector, directed, link |\n| `line.dashed` | Dashed Line | 160×0 | dotted, connector |\n| `rect` | Rectangle | 140×64 | box, card |\n| `square` | Square | 90×90 |  |\n| `ellipse` | Ellipse | 140×64 | oval |\n| `circle` | Circle | 90×90 |  |\n| `diamond` | Diamond | 140×64 | rhombus |\n| `triangle` | Triangle | 140×64 |  |\n| `hexagon` | Hexagon | 140×64 |  |\n| `pentagon` | Pentagon | 140×64 |  |\n| `octagon` | Octagon | 140×64 |  |\n| `parallelogram` | Parallelogram | 140×64 |  |\n| `trapezoid` | Trapezoid | 140×64 |  |\n| `cylinder` | Cylinder | 140×64 | can, tube |\n| `cloud-shape` | Cloud | 140×64 |  |\n| `callout` | Callout | 140×64 | speech, bubble |\n| `text` | Text | 120×36 |  |\n| `sticky` | Sticky note | 140×100 | postit, comment |\n| `card` | Card | 140×64 |  |\n| `frame` | Frame / Container (container) | 220×160 | group, container, boundary |\n| `pool.horizontal` | Pool (horizontal) (swimlane) | 360×220 | swimlane, lane, container, subsystem |\n| `pool.vertical` | Pool (vertical) (swimlane) | 220×360 | swimlane, lane, container, subsystem |\n| `image` | Image | 140×64 | picture |\n\n## Flowchart (`flowchart`)\n\n| Type | Label | Default size | Aliases |\n| --- | --- | --- | --- |\n| `flow.process` | Process | 140×64 | step, action |\n| `flow.alt-process` | Alternate Process | 140×64 |  |\n| `flow.decision` | Decision | 140×100 | if, branch |\n| `flow.terminator` | Start / End | 130×56 | terminator, begin, stop |\n| `flow.data` | Data (I/O) | 140×64 | input, output, io |\n| `flow.document` | Document | 140×64 |  |\n| `flow.multi-document` | Multi-Document | 140×64 |  |\n| `flow.predefined` | Predefined Process | 140×64 | subroutine |\n| `flow.manual-input` | Manual Input | 140×64 |  |\n| `flow.manual-op` | Manual Operation | 140×64 |  |\n| `flow.preparation` | Preparation | 140×64 |  |\n| `flow.stored-data` | Stored Data | 140×64 |  |\n| `flow.database` | Database | 140×64 | db, store |\n| `flow.display` | Display | 140×64 |  |\n| `flow.delay` | Delay | 140×64 |  |\n| `flow.merge` | Merge | 140×64 |  |\n| `flow.or` | Or | 140×64 |  |\n| `flow.connector` | On-page Connector | 48×48 |  |\n| `flow.off-page` | Off-page Connector | 140×64 |  |\n| `flow.annotation` | Annotation | 140×64 |  |\n\n## UML (`uml`)\n\n| Type | Label | Default size | Aliases |\n| --- | --- | --- | --- |\n| `uml.class` | Class | 140×64 | entity |\n| `uml.abstract` | Abstract Class | 140×64 |  |\n| `uml.interface` | Interface | 140×64 |  |\n| `uml.enum` | Enum | 140×64 |  |\n| `uml.object` | Object | 140×64 |  |\n| `uml.package` | Package | 140×64 |  |\n| `uml.component` | Component | 140×64 |  |\n| `uml.node` | Node / Deployment | 140×64 |  |\n| `uml.actor` | Actor | 60×100 | user, role |\n| `uml.usecase` | Use Case | 140×64 |  |\n| `uml.boundary` | Boundary | 140×64 |  |\n| `uml.control` | Control | 140×64 |  |\n| `uml.entity` | Entity | 140×64 |  |\n| `uml.state` | State | 140×64 |  |\n| `uml.initial` | Initial State | 36×36 |  |\n| `uml.final` | Final State | 40×40 |  |\n| `uml.decision` | Decision | 60×60 |  |\n| `uml.note` | Note / Comment | 140×64 | comment |\n\n## ERD (`erd`)\n\n| Type | Label | Default size | Aliases |\n| --- | --- | --- | --- |\n| `erd.entity` | Entity | 140×64 | table |\n| `erd.weak-entity` | Weak Entity | 140×64 |  |\n| `erd.attribute` | Attribute | 140×64 |  |\n| `erd.key-attribute` | Key Attribute | 140×64 | pk, primary key |\n| `erd.multivalued` | Multivalued Attribute | 140×64 |  |\n| `erd.derived` | Derived Attribute | 140×64 |  |\n| `erd.relationship` | Relationship | 140×64 |  |\n| `erd.identifying` | Identifying Relationship | 140×64 |  |\n\n## Architecture (`networking`)\n\n| Type | Label | Default size | Aliases |\n| --- | --- | --- | --- |\n| `net.user` | User / Client | 60×100 | client, person |\n| `net.admin` | Admin | 60×100 |  |\n| `net.browser` | Browser | 140×64 |  |\n| `net.mobile` | Mobile Device | 90×130 | phone |\n| `net.desktop` | Desktop | 140×64 |  |\n| `net.iot` | IoT Device | 140×64 |  |\n| `net.server` | Server | 140×64 |  |\n| `net.service` | Service | 140×64 | microservice |\n| `net.monolith` | Monolith | 140×64 |  |\n| `net.api-gateway` | API Gateway | 140×64 | api, gateway |\n| `net.load-balancer` | Load Balancer | 140×64 | lb |\n| `net.reverse-proxy` | Reverse Proxy | 140×64 | proxy |\n| `net.firewall` | Firewall | 140×64 | waf |\n| `net.router` | Router | 140×64 |  |\n| `net.switch` | Switch | 140×64 |  |\n| `net.dns` | DNS | 140×64 |  |\n| `net.cdn` | CDN | 140×64 | edge |\n| `net.cache` | Cache | 140×64 |  |\n| `net.queue` | Queue | 140×64 | mq |\n| `net.websocket` | Web Socket | 140×64 | ws |\n| `net.rest` | REST endpoint | 140×64 | http, api |\n| `net.internet` | Internet / Cloud | 140×64 | wan |\n| `net.datacenter` | Datacenter | 140×64 |  |\n| `net.container` | Container | 140×64 |  |\n| `net.pod` | Pod | 140×64 |  |\n| `net.vm` | Virtual Machine | 140×64 | vm |\n\n## Databases (`database`)\n\n| Type | Label | Default size | Aliases |\n| --- | --- | --- | --- |\n| `db.postgres` | PostgreSQL | 176×56 | postgres, pg |\n| `db.mysql` | MySQL | 176×56 |  |\n| `db.mariadb` | MariaDB | 176×56 |  |\n| `db.mongodb` | MongoDB | 176×56 | mongo |\n| `db.redis` | Redis | 176×56 |  |\n| `db.sqlite` | SQLite | 176×56 |  |\n| `db.mssql` | MS SQL Server | 176×56 | sqlserver, mssql |\n| `db.oracle` | Oracle | 176×56 |  |\n| `db.cassandra` | Cassandra | 176×56 |  |\n| `db.elasticsearch` | Elasticsearch | 176×56 | elastic, es |\n| `db.neo4j` | Neo4j | 176×56 | graph |\n| `db.couchdb` | CouchDB | 176×56 | couch |\n| `db.couchbase` | Couchbase | 176×56 |  |\n| `db.clickhouse` | ClickHouse | 176×56 |  |\n| `db.cockroachdb` | CockroachDB | 176×56 | cockroach |\n| `db.influxdb` | InfluxDB | 176×56 | influx, tsdb |\n| `db.snowflake` | Snowflake | 176×56 |  |\n| `db.supabase` | Supabase | 176×56 |  |\n\n## Messaging (`messaging`)\n\n| Type | Label | Default size | Aliases |\n| --- | --- | --- | --- |\n| `mq.kafka` | Apache Kafka | 176×56 | kafka |\n| `mq.rabbitmq` | RabbitMQ | 176×56 | rabbit, amqp |\n| `mq.nats` | NATS | 176×56 |  |\n| `mq.pulsar` | Apache Pulsar | 176×56 | pulsar |\n| `mq.mqtt` | MQTT | 176×56 |  |\n| `mq.pubsub` | Pub/Sub | 176×56 | pubsub |\n| `mq.activemq` | ActiveMQ | 176×56 | jms |\n\n## Languages (`language`)\n\n| Type | Label | Default size | Aliases |\n| --- | --- | --- | --- |\n| `lang.node` | Node.js | 176×56 | node, js runtime |\n| `lang.python` | Python | 176×56 | py |\n| `lang.java` | Java | 176×56 | jvm |\n| `lang.go` | Go | 176×56 | golang |\n| `lang.rust` | Rust | 176×56 |  |\n| `lang.dotnet` | C# / .NET | 176×56 | csharp, c# |\n| `lang.typescript` | TypeScript | 176×56 | ts |\n| `lang.javascript` | JavaScript | 176×56 | js |\n| `lang.ruby` | Ruby | 176×56 |  |\n| `lang.php` | PHP | 176×56 |  |\n| `lang.kotlin` | Kotlin | 176×56 |  |\n| `lang.swift` | Swift | 176×56 |  |\n| `lang.scala` | Scala | 176×56 |  |\n| `lang.elixir` | Elixir | 176×56 |  |\n| `lang.cpp` | C++ | 176×56 | cpp |\n\n## Frameworks (`framework`)\n\n| Type | Label | Default size | Aliases |\n| --- | --- | --- | --- |\n| `fw.express` | Express | 176×56 |  |\n| `fw.nestjs` | NestJS | 176×56 | nest |\n| `fw.spring` | Spring Boot | 176×56 | spring |\n| `fw.django` | Django | 176×56 |  |\n| `fw.flask` | Flask | 176×56 |  |\n| `fw.fastapi` | FastAPI | 176×56 |  |\n| `fw.rails` | Rails | 176×56 | ruby on rails |\n| `fw.laravel` | Laravel | 176×56 |  |\n| `fw.dotnet` | .NET | 176×56 |  |\n| `fw.react` | React | 176×56 |  |\n| `fw.vue` | Vue | 176×56 | vuejs |\n| `fw.angular` | Angular | 176×56 |  |\n| `fw.next` | Next.js | 176×56 | nextjs |\n| `fw.nuxt` | Nuxt | 176×56 | nuxtjs |\n| `fw.svelte` | Svelte | 176×56 | sveltekit |\n| `fw.electron` | Electron | 176×56 |  |\n| `fw.tailwind` | Tailwind | 176×56 | tailwindcss |\n| `fw.graphql` | GraphQL | 176×56 |  |\n\n## DevOps / CI-CD (`devops`)\n\n| Type | Label | Default size | Aliases |\n| --- | --- | --- | --- |\n| `ops.docker` | Docker | 176×56 |  |\n| `ops.k8s` | Kubernetes | 176×56 | k8s |\n| `ops.nginx` | Nginx | 176×56 |  |\n| `ops.haproxy` | HAProxy | 176×56 | lb |\n| `ops.traefik` | Traefik | 176×56 | traefik |\n| `ops.terraform` | Terraform | 176×56 | iac |\n| `ops.ansible` | Ansible | 176×56 |  |\n| `ops.jenkins` | Jenkins | 176×56 | ci |\n| `ops.gha` | GitHub Actions | 176×56 | ci, actions |\n| `ops.gitlab` | GitLab CI | 176×56 | ci |\n| `ops.argocd` | ArgoCD | 176×56 | gitops |\n| `ops.helm` | Helm | 176×56 |  |\n| `ops.istio` | Istio | 176×56 | service mesh |\n| `ops.envoy` | Envoy | 176×56 | proxy |\n| `ops.consul` | Consul | 176×56 |  |\n| `ops.vault` | Vault | 176×56 | secrets |\n| `ops.pm2` | PM2 | 176×56 |  |\n| `ops.maven` | Maven | 176×56 | build |\n\n## Auth / Identity (`auth`)\n\n| Type | Label | Default size | Aliases |\n| --- | --- | --- | --- |\n| `auth.oidc` | OpenID Connect | 176×56 | oidc |\n| `auth.jwt` | JWT | 176×56 | token |\n| `auth.keycloak` | Keycloak | 176×56 |  |\n| `auth.auth0` | Auth0 | 176×56 |  |\n| `auth.okta` | Okta | 176×56 |  |\n| `auth.firebase` | Firebase Auth | 176×56 | firebase |\n| `auth.oauth` | OAuth 2.0 | 176×56 | oauth |\n| `auth.saml` | SAML | 176×56 |  |\n| `auth.ldap` | LDAP | 176×56 |  |\n\n## Observability (`observability`)\n\n| Type | Label | Default size | Aliases |\n| --- | --- | --- | --- |\n| `obs.prometheus` | Prometheus | 176×56 | metrics |\n| `obs.grafana` | Grafana | 176×56 | dashboard |\n| `obs.datadog` | Datadog | 176×56 |  |\n| `obs.sentry` | Sentry | 176×56 | errors |\n| `obs.elasticsearch` | Elasticsearch | 176×56 | elk |\n| `obs.logstash` | Logstash | 176×56 | elk |\n| `obs.kibana` | Kibana | 176×56 | elk |\n| `obs.jaeger` | Jaeger | 176×56 | tracing |\n| `obs.otel` | OpenTelemetry | 176×56 | otel, tracing |\n| `obs.newrelic` | New Relic | 176×56 |  |\n\n## AWS (`aws`)\n\n| Type | Label | Default size | Aliases |\n| --- | --- | --- | --- |\n| `aws.ec2` | EC2 | 176×56 | aws |\n| `aws.lambda` | Lambda | 176×56 | aws |\n| `aws.s3` | S3 | 176×56 | aws |\n| `aws.ebs` | EBS | 176×56 | aws |\n| `aws.efs` | EFS | 176×56 | aws |\n| `aws.rds` | RDS | 176×56 | aws |\n| `aws.aurora` | Aurora | 176×56 | aws |\n| `aws.dynamodb` | DynamoDB | 176×56 | aws |\n| `aws.elasticache` | ElastiCache | 176×56 | aws |\n| `aws.redshift` | Redshift | 176×56 | aws |\n| `aws.vpc` | VPC | 176×56 | aws |\n| `aws.subnet` | Subnet | 176×56 | aws |\n| `aws.internet-gateway` | Internet Gateway | 176×56 | aws |\n| `aws.nat-gateway` | NAT Gateway | 176×56 | aws |\n| `aws.elb` | ELB | 176×56 | aws |\n| `aws.alb` | ALB | 176×56 | aws |\n| `aws.nlb` | NLB | 176×56 | aws |\n| `aws.api-gateway` | API Gateway | 176×56 | aws |\n| `aws.cloudfront` | CloudFront | 176×56 | aws |\n| `aws.route-53` | Route 53 | 176×56 | aws |\n| `aws.sqs` | SQS | 176×56 | aws |\n| `aws.sns` | SNS | 176×56 | aws |\n| `aws.eventbridge` | EventBridge | 176×56 | aws |\n| `aws.kinesis` | Kinesis | 176×56 | aws |\n| `aws.step-functions` | Step Functions | 176×56 | aws |\n| `aws.ecs` | ECS | 176×56 | aws |\n| `aws.eks` | EKS | 176×56 | aws |\n| `aws.fargate` | Fargate | 176×56 | aws |\n| `aws.ecr` | ECR | 176×56 | aws |\n| `aws.iam` | IAM | 176×56 | aws |\n| `aws.cognito` | Cognito | 176×56 | aws |\n| `aws.secrets-manager` | Secrets Manager | 176×56 | aws |\n| `aws.kms` | KMS | 176×56 | aws |\n| `aws.cloudwatch` | CloudWatch | 176×56 | aws |\n| `aws.cloudtrail` | CloudTrail | 176×56 | aws |\n| `aws.glue` | Glue | 176×56 | aws |\n| `aws.athena` | Athena | 176×56 | aws |\n| `aws.sagemaker` | SageMaker | 176×56 | aws |\n| `aws.appsync` | AppSync | 176×56 | aws |\n| `aws.amplify` | Amplify | 176×56 | aws |\n\n## Azure (`azure`)\n\n| Type | Label | Default size | Aliases |\n| --- | --- | --- | --- |\n| `azure.virtual-machine` | Virtual Machine | 176×56 | azure |\n| `azure.vm-scale-set` | VM Scale Set | 176×56 | azure |\n| `azure.app-service` | App Service | 176×56 | azure |\n| `azure.functions` | Functions | 176×56 | azure |\n| `azure.container-instances` | Container Instances | 176×56 | azure |\n| `azure.aks` | AKS | 176×56 | azure |\n| `azure.container-registry` | Container Registry | 176×56 | azure |\n| `azure.blob-storage` | Blob Storage | 176×56 | azure |\n| `azure.files` | Files | 176×56 | azure |\n| `azure.disk` | Disk | 176×56 | azure |\n| `azure.sql-database` | SQL Database | 176×56 | azure |\n| `azure.cosmos-db` | Cosmos DB | 176×56 | azure |\n| `azure.cache-for-redis` | Cache for Redis | 176×56 | azure |\n| `azure.synapse` | Synapse | 176×56 | azure |\n| `azure.virtual-network` | Virtual Network | 176×56 | azure |\n| `azure.load-balancer` | Load Balancer | 176×56 | azure |\n| `azure.application-gateway` | Application Gateway | 176×56 | azure |\n| `azure.front-door` | Front Door | 176×56 | azure |\n| `azure.service-bus` | Service Bus | 176×56 | azure |\n| `azure.event-hub` | Event Hub | 176×56 | azure |\n| `azure.event-grid` | Event Grid | 176×56 | azure |\n| `azure.key-vault` | Key Vault | 176×56 | azure |\n| `azure.entra-id` | Entra ID | 176×56 | azure |\n| `azure.monitor` | Monitor | 176×56 | azure |\n| `azure.application-insights` | Application Insights | 176×56 | azure |\n| `azure.logic-apps` | Logic Apps | 176×56 | azure |\n| `azure.api-management` | API Management | 176×56 | azure |\n\n## GCP (`gcp`)\n\n| Type | Label | Default size | Aliases |\n| --- | --- | --- | --- |\n| `gcp.compute-engine` | Compute Engine | 176×56 | gcp, google |\n| `gcp.app-engine` | App Engine | 176×56 | gcp, google |\n| `gcp.cloud-functions` | Cloud Functions | 176×56 | gcp, google |\n| `gcp.cloud-run` | Cloud Run | 176×56 | gcp, google |\n| `gcp.gke` | GKE | 176×56 | gcp, google |\n| `gcp.artifact-registry` | Artifact Registry | 176×56 | gcp, google |\n| `gcp.cloud-storage` | Cloud Storage | 176×56 | gcp, google |\n| `gcp.persistent-disk` | Persistent Disk | 176×56 | gcp, google |\n| `gcp.cloud-sql` | Cloud SQL | 176×56 | gcp, google |\n| `gcp.cloud-spanner` | Cloud Spanner | 176×56 | gcp, google |\n| `gcp.firestore` | Firestore | 176×56 | gcp, google |\n| `gcp.bigtable` | Bigtable | 176×56 | gcp, google |\n| `gcp.bigquery` | BigQuery | 176×56 | gcp, google |\n| `gcp.memorystore` | Memorystore | 176×56 | gcp, google |\n| `gcp.vpc` | VPC | 176×56 | gcp, google |\n| `gcp.cloud-load-balancing` | Cloud Load Balancing | 176×56 | gcp, google |\n| `gcp.cloud-cdn` | Cloud CDN | 176×56 | gcp, google |\n| `gcp.cloud-dns` | Cloud DNS | 176×56 | gcp, google |\n| `gcp.pub-sub` | Pub/Sub | 176×56 | gcp, google |\n| `gcp.dataflow` | Dataflow | 176×56 | gcp, google |\n| `gcp.cloud-tasks` | Cloud Tasks | 176×56 | gcp, google |\n| `gcp.iam` | IAM | 176×56 | gcp, google |\n| `gcp.secret-manager` | Secret Manager | 176×56 | gcp, google |\n| `gcp.cloud-kms` | Cloud KMS | 176×56 | gcp, google |\n| `gcp.cloud-monitoring` | Cloud Monitoring | 176×56 | gcp, google |\n| `gcp.vertex-ai` | Vertex AI | 176×56 | gcp, google |\n",

package/dist-cjs/resources/index.js

@@ -8,10 +8,12 @@ const requestShape_js_1 = require("../domain/requestShape.js");
 const summarize_js_2 = require("../domain/scenario/summarize.js");
 const truncate_js_1 = require("../lib/truncate.js");
 const lookups_js_1 = require("../tools/lookups.js");
+const uiTestShared_js_1 = require("../tools/uiTestShared.js");
 const contracts_js_1 = require("./docs/contracts.js");
 const diagramSchema_js_1 = require("./docs/diagramSchema.js");
 const overview_js_1 = require("./docs/overview.js");
 const requestSchema_js_1 = require("./docs/requestSchema.js");
+const uiTests_js_1 = require("./docs/uiTests.js");
 const aiPacks_js_1 = require("./generated/aiPacks.js");
 // Resources: reference material a model reads on demand (schemas, the
 // scripting API, the scenario node reference, the diagram component catalog
@@ -43,6 +45,8 @@ const registerResources = (server, ctx) => {
         { name: "docs-scripting-db-queries", uri: "callman://docs/scripting/db-queries", title: "Scripting: database queries", description: "How pm.db.* queries work per database provider.", body: () => (0, callman_core_1.renderDbQueryReference)() },
         { name: "docs-scripting-templating", uri: "callman://docs/scripting/templating", title: "Templating: {{variables}} and generators", description: "Variable resolution order and dynamic generators ({{$uuid}}, {{$randomInt}} …).", body: () => aiPacks_js_1.AI_PACKS.templating ?? "" },
         { name: "docs-scripting-sandbox", uri: "callman://docs/scripting/sandbox", title: "Scripting: sandbox and feature policy", description: "Globals available in the script sandbox and the feature policy for DB/Kafka capabilities.", body: () => `${aiPacks_js_1.AI_PACKS["sandbox-globals"] ?? ""}\n\n${aiPacks_js_1.AI_PACKS["feature-policy"] ?? ""}` },
+        { name: "docs-ui-tests", uri: "callman://docs/ui-tests", title: "UI tests", description: "UI-test concepts: platforms, header, preconditions and sub-flows, datasets and DB sources, hybrid data steps, settings, where runs happen (server runner vs your desktop app), reports and screenshots, schedules, review, masked secrets, and the recommended workflow.", body: () => uiTests_js_1.UI_TESTS_DOC },
+        { name: "docs-ui-test-steps", uri: "callman://docs/ui-test-steps", title: "UI-test step reference", description: "Every UI-test step (web / Android / iOS / hybrid data steps): aliases, platforms, forms, arguments and an example — generated from callman-core, the same catalog the linter uses.", body: () => (0, callman_core_1.renderUiTestCommandReference)() },
         { name: "docs-diagram-schema", uri: "callman://docs/diagram-schema", title: "Diagram schema", description: "Every node and edge field, styling (colors, borders, fonts), containers, edge semantics and colors, patch operations and merge rules, and the stored DiagramDocument.", body: () => diagramSchema_js_1.DIAGRAM_SCHEMA_DOC },
     ];
     for (const doc of staticDocs) {
@@ -105,6 +109,13 @@ const registerResources = (server, ctx) => {
         const resolved = await (0, lookups_js_1.resolveScenario)(ctx, ws?.id ?? null, variable(variables.scenario));
         return json(uri, { scenario: (0, summarize_js_2.scenarioFullView)(resolved.item) });
     });
+    server.registerResource("ui-test", new mcp_js_1.ResourceTemplate("callman://ui-test/{flow}", { list: undefined }), { title: "A UI test", description: "One UI test (id or name): header, settings, dataset (masked), preconditions, numbered steps, lint result and YAML.", mimeType: "application/json" }, async (uri, variables) => {
+        const resolved = await (0, lookups_js_1.resolveUiTest)(ctx, null, variable(variables.flow));
+        const fresh = await ctx.api.getUiTest(resolved.item.clientId);
+        const context = await (0, uiTestShared_js_1.lintContext)(ctx, fresh.workspaceId);
+        const lint = (0, uiTestShared_js_1.lintFlow)(fresh.content ?? "", fresh.platform, context, [...Object.keys(fresh.dataset ?? {}), ...Object.keys(fresh.datasetSources ?? {})]);
+        return json(uri, { uiTest: await (0, uiTestShared_js_1.flowDetailView)(ctx, fresh, lint, context, { workspaces: await ctx.listWorkspaces() }) });
+    });
     server.registerResource("diagram", new mcp_js_1.ResourceTemplate("callman://diagram/{diagram}", { list: undefined }), { title: "A diagram", description: "One diagram (id or name) as nodes/edges with edge semantics.", mimeType: "application/json" }, async (uri, variables) => {
         const ref = variable(variables.diagram);
         const resolved = await (0, lookups_js_1.resolveDiagram)(ctx, ref);

package/dist-cjs/server.js

@@ -7,8 +7,8 @@ const index_js_1 = require("./prompts/index.js");
 const index_js_2 = require("./resources/index.js");
 const index_js_3 = require("./tools/index.js");
 exports.SERVER_NAME = "callman";
-exports.SERVER_INSTRUCTIONS = `Callman is an API client and test platform: collections of HTTP requests, environments with variables, multi-step test scenarios, and system-design diagrams. You act as the user whose token is configured, within that token's permissions and the user's workspace role.
-Start with whoami. The token works in every workspace the user belongs to: lists and searches cover all of them unless you pass \`workspace\`, names are looked up in the default workspace first and then everywhere, and results say which workspace they come from; creating something needs a target (\`workspace\`, the collection it goes into, or the session default — use_workspace). Names work as references (collection, folder, request, environment, scenario, diagram, workspace); ambiguous names return candidates. List before creating — duplicate collection names are refused. Preview large writes with dryRun: true; deletions need confirm: true and the user's agreement. Environment values are masked unless the token holds environments:reveal and you pass reveal: true; never write a masked value back. Arguments are strict — an unknown key is an error listing the valid ones. Read callman://docs/overview for the conventions, callman://docs/request-schema before creating or editing requests, callman://docs/contracts before contract checks, callman://docs/scenario-nodes before scenarios, and callman://docs/diagram-schema + callman://diagram/components before diagrams.`;
+exports.SERVER_INSTRUCTIONS = `Callman is an API client and test platform: collections of HTTP requests, environments with variables, multi-step API test scenarios, UI tests (web / Android / iOS flows), and system-design diagrams. You act as the user whose token is configured, within that token's permissions and the user's workspace role.
+Start with whoami. The token works in every workspace the user belongs to: lists and searches cover all of them unless you pass \`workspace\`, names are looked up in the default workspace first and then everywhere, and results say which workspace they come from; creating something needs a target (\`workspace\`, the collection it goes into, or the session default — use_workspace). Names work as references (collection, folder, request, environment, scenario, UI test, diagram, workspace); ambiguous names return candidates. List before creating — duplicate collection names are refused. Preview large writes with dryRun: true; deletions need confirm: true and the user's agreement. Environment values are masked unless the token holds environments:reveal and you pass reveal: true; never write a masked value back. Arguments are strict — an unknown key is an error listing the valid ones. Read callman://docs/overview for the conventions, callman://docs/request-schema before creating or editing requests, callman://docs/contracts before contract checks, callman://docs/scenario-nodes before scenarios, callman://docs/ui-tests (+ callman://docs/ui-test-steps) before UI tests — they run on the server runner (web) or the user's own desktop app (list_ui_test_devices), and callman://docs/diagram-schema + callman://diagram/components before diagrams.`;
 const createCallmanMcpServer = (options) => {
     const version = options.version ?? "0.0.0";
     const context = (0, context_js_1.createToolContext)({

package/dist-cjs/tools/catalog.js

@@ -65,6 +65,37 @@ exports.TOOL_CATALOG = [
     { name: "submit_scenario", group: "scenarios", scopes: ["scenarios:write"] },
     { name: "delete_scenario", group: "scenarios", scopes: ["scenarios:delete"], destructive: true },
     { name: "run_scenario", group: "scenarios", scopes: ["scenarios:run"] },
+    // UI tests (web / Android / iOS flows): ui_tests:* scopes. Runs start on the
+    // server runner (web) or the user's desktop app (desktop relay).
+    { name: "list_ui_tests", group: "uiTests", scopes: ["ui_tests:read"] },
+    { name: "get_ui_test", group: "uiTests", scopes: ["ui_tests:read"] },
+    { name: "validate_ui_test", group: "uiTests", scopes: ["ui_tests:read"] },
+    { name: "create_ui_test", group: "uiTests", scopes: ["ui_tests:write"] },
+    { name: "update_ui_test", group: "uiTests", scopes: ["ui_tests:write"] },
+    { name: "patch_ui_test_steps", group: "uiTests", scopes: ["ui_tests:write"] },
+    { name: "delete_ui_test", group: "uiTests", scopes: ["ui_tests:delete"], destructive: true },
+    { name: "list_ui_test_folders", group: "uiTests", scopes: ["ui_tests:read"] },
+    { name: "create_ui_test_folder", group: "uiTests", scopes: ["ui_tests:write"] },
+    { name: "update_ui_test_folder", group: "uiTests", scopes: ["ui_tests:write"] },
+    { name: "delete_ui_test_folder", group: "uiTests", scopes: ["ui_tests:delete"], destructive: true },
+    { name: "move_ui_tests", group: "uiTests", scopes: ["ui_tests:write"] },
+    { name: "list_ui_test_devices", group: "uiTests", scopes: ["ui_tests:read"] },
+    { name: "run_ui_test", group: "uiTests", scopes: ["ui_tests:run"] },
+    { name: "run_ui_tests", group: "uiTests", scopes: ["ui_tests:run"] },
+    { name: "get_ui_test_run", group: "uiTests", scopes: ["ui_tests:read"] },
+    { name: "list_ui_test_runs", group: "uiTests", scopes: ["ui_tests:read"] },
+    { name: "cancel_ui_test_run", group: "uiTests", scopes: ["ui_tests:run"] },
+    { name: "list_ui_test_schedules", group: "uiTests", scopes: ["ui_tests:read"] },
+    { name: "create_ui_test_schedule", group: "uiTests", scopes: ["ui_tests:write"] },
+    { name: "update_ui_test_schedule", group: "uiTests", scopes: ["ui_tests:write"] },
+    { name: "delete_ui_test_schedule", group: "uiTests", scopes: ["ui_tests:delete"], destructive: true },
+    { name: "run_ui_test_schedule_now", group: "uiTests", scopes: ["ui_tests:run"] },
+    { name: "submit_ui_test", group: "uiTests", scopes: ["ui_tests:write"] },
+    { name: "review_ui_test", group: "uiTests", scopes: ["ui_tests:write"] },
+    { name: "list_ui_test_versions", group: "uiTests", scopes: ["ui_tests:read"] },
+    { name: "restore_ui_test_version", group: "uiTests", scopes: ["ui_tests:write"] },
+    { name: "export_ui_tests", group: "uiTests", scopes: ["ui_tests:read"] },
+    { name: "import_ui_tests", group: "uiTests", scopes: ["ui_tests:write"] },
     { name: "list_diagrams", group: "diagrams", scopes: ["diagrams:read"] },
     { name: "get_diagram", group: "diagrams", scopes: ["diagrams:read"] },
     { name: "create_diagram", group: "diagrams", scopes: ["diagrams:write"] },

package/dist-cjs/tools/index.js

@@ -11,6 +11,9 @@ const meta_js_1 = require("./meta.js");
 const requests_js_1 = require("./requests.js");
 const run_js_1 = require("./run.js");
 const scenarios_js_1 = require("./scenarios.js");
+const uiTestAdmin_js_1 = require("./uiTestAdmin.js");
+const uiTestRuns_js_1 = require("./uiTestRuns.js");
+const uiTests_js_1 = require("./uiTests.js");
 const workspaces_js_1 = require("./workspaces.js");
 var catalog_js_1 = require("./catalog.js");
 Object.defineProperty(exports, "TOOL_CATALOG", { enumerable: true, get: function () { return catalog_js_1.TOOL_CATALOG; } });
@@ -43,6 +46,9 @@ const registerAllTools = (host, ctx) => {
     (0, import_js_1.registerImportTools)(server, ctx);
     (0, environments_js_1.registerEnvironmentTools)(server, ctx);
     (0, scenarios_js_1.registerScenarioTools)(server, ctx);
+    (0, uiTests_js_1.registerUiTestTools)(server, ctx);
+    (0, uiTestRuns_js_1.registerUiTestRunTools)(server, ctx);
+    (0, uiTestAdmin_js_1.registerUiTestAdminTools)(server, ctx);
     (0, diagrams_js_1.registerDiagramTools)(server, ctx);
     if (ctx.config.allowLocalRun)
         (0, run_js_1.registerRunTools)(server, ctx);

package/dist-cjs/tools/lookups.js

@@ -1,6 +1,6 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
-exports.sameName = exports.connectionResolverFor = exports.loadConnections = exports.resolveDiagram = exports.loadDiagrams = exports.resolveScenarioFolder = exports.resolveScenario = exports.loadScenarioCatalog = exports.resolveEnvironment = exports.loadEnvironments = exports.resolveRequest = exports.searchRequests = exports.requestIdsInFolder = exports.resolveFolder = exports.loadTree = exports.resolveCollection = exports.findEnvironment = exports.findRequest = exports.findCollection = exports.findInWorkspaces = exports.workspaceTag = exports.workspacesForListing = exports.isAllWorkspaces = exports.loadCollections = void 0;
+exports.sameName = exports.connectionResolverFor = exports.loadConnections = exports.resolveUiTest = exports.resolveUiTestFolder = exports.uiTestFolderSubtree = exports.loadUiTestFolders = exports.loadUiTests = exports.resolveDiagram = exports.loadDiagrams = exports.resolveScenarioFolder = exports.resolveScenario = exports.loadScenarioCatalog = exports.resolveEnvironment = exports.loadEnvironments = exports.resolveRequest = exports.searchRequests = exports.requestIdsInFolder = exports.resolveFolder = exports.loadTree = exports.resolveCollection = exports.findEnvironment = exports.findRequest = exports.findCollection = exports.findInWorkspaces = exports.workspaceTag = exports.workspacesForListing = exports.isAllWorkspaces = exports.loadCollections = void 0;
 const errors_js_1 = require("../client/errors.js");
 const requestIntegrations_js_1 = require("../domain/requestIntegrations.js");
 const resolve_js_1 = require("../domain/resolve.js");
@@ -251,6 +251,88 @@ const resolveDiagram = async (ctx, ref, options = {}) => {
     return { ...resolved, item: await ctx.api.getDiagram(resolved.item.id) };
 };
 exports.resolveDiagram = resolveDiagram;
+const loadUiTests = (ctx) => ctx.cache.getOrSet("uitests:list", () => ctx.api.listUiTests(true));
+exports.loadUiTests = loadUiTests;
+const loadUiTestFolders = (ctx) => ctx.cache.getOrSet("uitests:folders", async () => {
+    const raw = await ctx.api.listUiTestFolders();
+    const byId = new Map(raw.map((folder) => [folder.clientId, folder]));
+    const pathOf = (folder) => {
+        const names = [folder.name];
+        const seen = new Set([folder.clientId]);
+        let parent = folder.parentId ? byId.get(folder.parentId) : undefined;
+        while (parent && !seen.has(parent.clientId)) {
+            names.unshift(parent.name);
+            seen.add(parent.clientId);
+            parent = parent.parentId ? byId.get(parent.parentId) : undefined;
+        }
+        return { path: names.join("/"), depth: names.length - 1 };
+    };
+    return raw.map((folder) => ({
+        id: folder.clientId,
+        name: folder.name,
+        parentId: folder.parentId ?? null,
+        ...pathOf(folder),
+        workspaceId: folder.workspaceId,
+        groupId: folder.groupId ?? null,
+        sortOrder: folder.sortOrder ?? 0,
+    }));
+});
+exports.loadUiTestFolders = loadUiTestFolders;
+// Folder ids of `folderId` and every folder below it.
+const uiTestFolderSubtree = (folders, folderId) => {
+    const out = new Set([folderId]);
+    let grew = true;
+    while (grew) {
+        grew = false;
+        for (const folder of folders) {
+            if (folder.parentId && out.has(folder.parentId) && !out.has(folder.id)) {
+                out.add(folder.id);
+                grew = true;
+            }
+        }
+    }
+    return out;
+};
+exports.uiTestFolderSubtree = uiTestFolderSubtree;
+const workspaceFilter = async (ctx, workspace) => workspace && workspace.trim() && !(0, exports.isAllWorkspaces)(workspace) ? ctx.resolveWorkspace(workspace) : null;
+const resolveUiTestFolder = async (ctx, workspace, ref, options = {}) => {
+    const ws = await workspaceFilter(ctx, workspace);
+    const folders = (await (0, exports.loadUiTestFolders)(ctx)).filter((folder) => !ws || folder.workspaceId === ws.id);
+    return (0, resolve_js_1.resolveRef)("UI-test folder", ref, folders, { describe: (folder) => ({ path: folder.path }), aliases: (folder) => [folder.path], exact: options.exact });
+};
+exports.resolveUiTestFolder = resolveUiTestFolder;
+// A flow by clientId or name (unique across the workspaces in scope). The
+// cached list may miss a flow created a moment ago — an unknown ref is tried
+// as a clientId before giving up.
+const resolveUiTest = async (ctx, workspace, ref, options = {}) => {
+    const ws = await workspaceFilter(ctx, workspace);
+    const [flows, workspaces, folders] = await Promise.all([(0, exports.loadUiTests)(ctx), ctx.listWorkspaces(), (0, exports.loadUiTestFolders)(ctx).catch(() => [])]);
+    const workspaceName = new Map(workspaces.map((entry) => [entry.id, entry.name]));
+    const folderPath = new Map(folders.map((folder) => [folder.id, folder.path]));
+    const items = flows.filter((flow) => !ws || flow.workspaceId === ws.id).map((flow) => ({ ...flow, id: flow.clientId }));
+    try {
+        return (0, resolve_js_1.resolveRef)("UI test", ref, items, {
+            describe: (flow) => ({ platform: flow.platform, folder: flow.folderId ? (folderPath.get(flow.folderId) ?? null) : null, ...(workspaces.length > 1 ? { workspace: workspaceName.get(flow.workspaceId) ?? flow.workspaceId } : {}) }),
+            exact: options.exact,
+        });
+    }
+    catch (error) {
+        if (!(error instanceof errors_js_1.ToolInputError) || error.code !== "REF_NOT_FOUND" || /\s/.test(ref.trim()))
+            throw error;
+        try {
+            const fresh = await ctx.api.getUiTest(ref.trim());
+            if (ws && fresh.workspaceId !== ws.id)
+                throw error;
+            return { item: { ...fresh, id: fresh.clientId }, matched: "id" };
+        }
+        catch (lookup) {
+            if (lookup instanceof errors_js_1.CallmanApiError && (lookup.status === 404 || lookup.status === 400))
+                throw error;
+            throw lookup;
+        }
+    }
+};
+exports.resolveUiTest = resolveUiTest;
 const loadConnections = (ctx, ws) => ctx.cache.getOrSet(`connections:${ws}`, () => ctx.api.listConnections(ws));
 exports.loadConnections = loadConnections;
 // Connection resolver for request integration steps — loaded only when a

package/dist-cjs/tools/shared.js

@@ -1,6 +1,6 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
-exports.requireConfirm = exports.dryRunArg = exports.confirmArg = exports.cursorArgs = exports.refArg = exports.workspaceArg = exports.RUN = exports.DESTRUCTIVE = exports.IDEMPOTENT_WRITE = exports.WRITE = exports.READ = exports.guarded = exports.failureOf = exports.toolFailure = exports.ok = void 0;
+exports.requireConfirm = exports.dryRunArg = exports.confirmArg = exports.cursorArgs = exports.refArg = exports.workspaceArg = exports.RUN = exports.DESTRUCTIVE = exports.IDEMPOTENT_WRITE = exports.WRITE = exports.READ = exports.guarded = exports.failureOf = exports.toolFailure = exports.okWithImages = exports.ok = void 0;
 const zod_1 = require("zod");
 const errors_js_1 = require("../client/errors.js");
 const truncate_js_1 = require("../lib/truncate.js");
@@ -17,6 +17,17 @@ const ok = (structured, text) => ({
     structuredContent: structured,
 });
 exports.ok = ok;
+// `ok` plus MCP image blocks (screenshots): the structured result stays JSON,
+// each image follows as { type: "image", data: <base64>, mimeType } preceded
+// by a one-line caption so the model knows which is which.
+const okWithImages = (structured, images) => {
+    const result = (0, exports.ok)(structured);
+    for (const image of images) {
+        result.content.push({ type: "text", text: image.caption }, { type: "image", data: image.data, mimeType: image.mimeType });
+    }
+    return result;
+};
+exports.okWithImages = okWithImages;
 const toolFailure = (error) => {
     let payload;
     if (error instanceof errors_js_1.CallmanApiError) {