@joshski/dust 0.1.75 → 0.1.77

package/dist/agent-events.d.ts

@@ -30,6 +30,7 @@ export interface EventMessage {
     timestamp: string;
     sessionId: string;
     repository: string;
+    repoId?: number;
     agentSessionId?: string;
     event: AgentSessionEvent;
 }

package/dist/audits.js

@@ -719,6 +719,65 @@ function repositoryContext() {
     - [ ] A new agent reading only this document could make sensible high-level suggestions
   `;
 }
+function slowTests() {
+  return dedent`
+    # Slow Tests
+
+    Identify slow-running tests that impact feedback loop speed.
+
+    ${ideasHint}
+
+    ## Scope
+
+    Focus on these areas:
+
+    1. **Test execution times** - Identify tests that exceed a reasonable duration threshold (e.g., 100ms for unit tests, 1s for integration tests)
+    2. **I/O-bound tests** - Tests that perform actual file system, network, or database operations
+    3. **Sleep/delay usage** - Tests using \`setTimeout\`, \`sleep\`, or similar timing functions
+    4. **Missing mocks** - Tests that make real HTTP requests or database calls instead of using stubs
+    5. **Setup overhead** - Expensive \`beforeEach\`/\`beforeAll\` setup that could be optimized or shared
+    6. **Serial test execution** - Tests that could run in parallel but are forced to run serially
+
+    ## Analysis Steps
+
+    1. Run the test suite with timing output: \`npm test -- --reporter=verbose\` or equivalent
+    2. Identify tests taking longer than the threshold (100ms+ for unit, 1s+ for integration)
+    3. Search for \`setTimeout\`, \`sleep\`, \`delay\`, and similar timing patterns in test files
+    4. Look for real I/O: \`fetch\`, \`axios\`, database clients, file system operations without mocks
+    5. Review \`beforeEach\`/\`beforeAll\` blocks for expensive operations
+    6. Check test configuration for parallelization settings
+
+    ## Output
+
+    For each slow test identified, provide:
+    - **Test name** - The describe/it block name
+    - **File path** - Location of the test
+    - **Duration** - How long the test takes (if measurable)
+    - **Cause** - Why the test is slow (I/O, sleep, setup, etc.)
+    - **Suggestion** - Specific optimization (mock the API, use fake timers, share setup, etc.)
+
+    ## Principles
+
+    - [Fast Feedback Loops](../principles/fast-feedback-loops.md) - Tests should run quickly for tight iteration cycles
+    - [Fast Feedback](../principles/fast-feedback.md) - Slow tests discourage frequent validation
+    - [Keep Unit Tests Pure](../principles/keep-unit-tests-pure.md) - Pure tests are faster and more reliable
+    - [Stubs Over Mocks](../principles/stubs-over-mocks.md) - Use stubs to avoid slow real dependencies
+
+    ## Blocked By
+
+    (none)
+
+    ## Definition of Done
+
+    - [ ] Ran test suite with timing information
+    - [ ] Listed tests exceeding duration thresholds (100ms unit, 1s integration)
+    - [ ] Identified tests using sleep/setTimeout/delay patterns
+    - [ ] Found tests with unmocked I/O (network, database, file system)
+    - [ ] Reviewed beforeEach/beforeAll for optimization opportunities
+    - [ ] Checked test parallelization configuration
+    - [ ] Proposed ideas for optimizing the slowest tests
+  `;
+}
 function ubiquitousLanguage() {
   return dedent`
     # Ubiquitous Language
@@ -772,6 +831,75 @@ function ubiquitousLanguage() {
     - [ ] Proposed ideas for standardizing inconsistent terminology
   `;
 }
+function uxAudit() {
+  return dedent`
+    # UX Audit
+
+    Review the end user experience by capturing visual or interactive evidence at key scenarios.
+
+    ${ideasHint}
+
+    ## Scope
+
+    1. **Identify key scenarios** - What are the main user journeys? (e.g., signup, login, checkout, onboarding, core workflows)
+    2. **Capture evidence** - For each scenario:
+       - Web apps: Take screenshots at each step using browser automation (Playwright, Puppeteer, Cypress, or similar)
+       - Terminal apps: Capture command output and interactive sessions
+    3. **Review captured evidence** for UX issues:
+       - Confusing or unclear states
+       - Missing feedback or loading indicators
+       - Error messages that don't guide recovery
+       - Inconsistent styling or layout
+    4. **Document findings** with screenshots/output and specific recommendations
+
+    ## Applicability
+
+    Determine the application type and available tooling:
+    - If browser tests exist (Playwright, Puppeteer, Cypress), extend them to capture screenshots
+    - If no browser tests exist, write a standalone script for key scenarios
+    - For terminal apps, capture representative sessions using command output or terminal recording
+
+    If the project has no user-facing interface, document that finding and skip the detailed analysis.
+
+    ## Analysis Steps
+
+    1. Identify the application type (web, terminal, hybrid, no UI)
+    2. List the key user scenarios from documentation, tests, or code analysis
+    3. Capture screenshots or output at each stage of each scenario
+    4. Store artifacts in a temporary directory for review during this audit
+    5. Review each artifact for UX issues
+    6. Document findings with evidence and specific recommendations
+
+    ## Output
+
+    For each UX issue identified, provide:
+    - **Location** - Which scenario and step
+    - **Evidence** - Screenshot filename or captured output
+    - **Problem** - What's wrong from the user's perspective
+    - **Impact** - How it affects the user's ability to complete their goal
+    - **Recommendation** - Specific fix
+    - **Verification** - How to verify the fix (e.g., "Screenshot at step 3 should show success message instead of spinner")
+
+    ## Principles
+
+    - [Actionable Errors](../principles/actionable-errors.md) - Error messages should tell users what to do next
+    - [Unsurprising UX](../principles/unsurprising-ux.md) - The interface should be as guessable as possible
+
+    ## Blocked By
+
+    (none)
+
+    ## Definition of Done
+
+    - [ ] Identified the application type (web, terminal, hybrid, or no UI)
+    - [ ] Listed key user scenarios
+    - [ ] Captured screenshots or output at each stage of key scenarios
+    - [ ] Reviewed evidence for UX issues
+    - [ ] Documented findings with evidence and recommendations
+    - [ ] Included verification criteria for each issue
+    - [ ] Created ideas for any UX improvements needed
+  `;
+}
 var stockAuditFunctions = {
   "agent-developer-experience": agentDeveloperExperience,
   "component-reuse": componentReuse,
@@ -787,9 +915,11 @@ var stockAuditFunctions = {
   "refactoring-opportunities": refactoringOpportunities,
   "repository-context": repositoryContext,
   "security-review": securityReview,
+  "slow-tests": slowTests,
   "stale-ideas": staleIdeas,
   "test-coverage": testCoverage,
-  "ubiquitous-language": ubiquitousLanguage
+  "ubiquitous-language": ubiquitousLanguage,
+  "ux-audit": uxAudit
 };
 function loadStockAudits() {
   return Object.entries(stockAuditFunctions).sort(([a], [b]) => a.localeCompare(b)).map(([name, render]) => {

package/dist/bucket/repository-loop.d.ts

@@ -30,6 +30,7 @@ export declare function buildEventMessage(parameters: {
     sequence: number;
     sessionId: string;
     repository: string;
+    repoId?: number;
     event: AgentSessionEvent;
     agentSessionId?: string;
 }): EventMessage;

package/dist/dust.js

@@ -275,7 +275,7 @@ async function loadSettings(cwd, fileSystem) {
 }
 
 // lib/version.ts
-var DUST_VERSION = "0.1.75";
+var DUST_VERSION = "0.1.77";
 
 // lib/session.ts
 var DUST_UNATTENDED = "DUST_UNATTENDED";
@@ -1266,6 +1266,65 @@ function repositoryContext() {
     - [ ] A new agent reading only this document could make sensible high-level suggestions
   `;
 }
+function slowTests() {
+  return dedent`
+    # Slow Tests
+
+    Identify slow-running tests that impact feedback loop speed.
+
+    ${ideasHint}
+
+    ## Scope
+
+    Focus on these areas:
+
+    1. **Test execution times** - Identify tests that exceed a reasonable duration threshold (e.g., 100ms for unit tests, 1s for integration tests)
+    2. **I/O-bound tests** - Tests that perform actual file system, network, or database operations
+    3. **Sleep/delay usage** - Tests using \`setTimeout\`, \`sleep\`, or similar timing functions
+    4. **Missing mocks** - Tests that make real HTTP requests or database calls instead of using stubs
+    5. **Setup overhead** - Expensive \`beforeEach\`/\`beforeAll\` setup that could be optimized or shared
+    6. **Serial test execution** - Tests that could run in parallel but are forced to run serially
+
+    ## Analysis Steps
+
+    1. Run the test suite with timing output: \`npm test -- --reporter=verbose\` or equivalent
+    2. Identify tests taking longer than the threshold (100ms+ for unit, 1s+ for integration)
+    3. Search for \`setTimeout\`, \`sleep\`, \`delay\`, and similar timing patterns in test files
+    4. Look for real I/O: \`fetch\`, \`axios\`, database clients, file system operations without mocks
+    5. Review \`beforeEach\`/\`beforeAll\` blocks for expensive operations
+    6. Check test configuration for parallelization settings
+
+    ## Output
+
+    For each slow test identified, provide:
+    - **Test name** - The describe/it block name
+    - **File path** - Location of the test
+    - **Duration** - How long the test takes (if measurable)
+    - **Cause** - Why the test is slow (I/O, sleep, setup, etc.)
+    - **Suggestion** - Specific optimization (mock the API, use fake timers, share setup, etc.)
+
+    ## Principles
+
+    - [Fast Feedback Loops](../principles/fast-feedback-loops.md) - Tests should run quickly for tight iteration cycles
+    - [Fast Feedback](../principles/fast-feedback.md) - Slow tests discourage frequent validation
+    - [Keep Unit Tests Pure](../principles/keep-unit-tests-pure.md) - Pure tests are faster and more reliable
+    - [Stubs Over Mocks](../principles/stubs-over-mocks.md) - Use stubs to avoid slow real dependencies
+
+    ## Blocked By
+
+    (none)
+
+    ## Definition of Done
+
+    - [ ] Ran test suite with timing information
+    - [ ] Listed tests exceeding duration thresholds (100ms unit, 1s integration)
+    - [ ] Identified tests using sleep/setTimeout/delay patterns
+    - [ ] Found tests with unmocked I/O (network, database, file system)
+    - [ ] Reviewed beforeEach/beforeAll for optimization opportunities
+    - [ ] Checked test parallelization configuration
+    - [ ] Proposed ideas for optimizing the slowest tests
+  `;
+}
 function ubiquitousLanguage() {
   return dedent`
     # Ubiquitous Language
@@ -1319,6 +1378,75 @@ function ubiquitousLanguage() {
     - [ ] Proposed ideas for standardizing inconsistent terminology
   `;
 }
+function uxAudit() {
+  return dedent`
+    # UX Audit
+
+    Review the end user experience by capturing visual or interactive evidence at key scenarios.
+
+    ${ideasHint}
+
+    ## Scope
+
+    1. **Identify key scenarios** - What are the main user journeys? (e.g., signup, login, checkout, onboarding, core workflows)
+    2. **Capture evidence** - For each scenario:
+       - Web apps: Take screenshots at each step using browser automation (Playwright, Puppeteer, Cypress, or similar)
+       - Terminal apps: Capture command output and interactive sessions
+    3. **Review captured evidence** for UX issues:
+       - Confusing or unclear states
+       - Missing feedback or loading indicators
+       - Error messages that don't guide recovery
+       - Inconsistent styling or layout
+    4. **Document findings** with screenshots/output and specific recommendations
+
+    ## Applicability
+
+    Determine the application type and available tooling:
+    - If browser tests exist (Playwright, Puppeteer, Cypress), extend them to capture screenshots
+    - If no browser tests exist, write a standalone script for key scenarios
+    - For terminal apps, capture representative sessions using command output or terminal recording
+
+    If the project has no user-facing interface, document that finding and skip the detailed analysis.
+
+    ## Analysis Steps
+
+    1. Identify the application type (web, terminal, hybrid, no UI)
+    2. List the key user scenarios from documentation, tests, or code analysis
+    3. Capture screenshots or output at each stage of each scenario
+    4. Store artifacts in a temporary directory for review during this audit
+    5. Review each artifact for UX issues
+    6. Document findings with evidence and specific recommendations
+
+    ## Output
+
+    For each UX issue identified, provide:
+    - **Location** - Which scenario and step
+    - **Evidence** - Screenshot filename or captured output
+    - **Problem** - What's wrong from the user's perspective
+    - **Impact** - How it affects the user's ability to complete their goal
+    - **Recommendation** - Specific fix
+    - **Verification** - How to verify the fix (e.g., "Screenshot at step 3 should show success message instead of spinner")
+
+    ## Principles
+
+    - [Actionable Errors](../principles/actionable-errors.md) - Error messages should tell users what to do next
+    - [Unsurprising UX](../principles/unsurprising-ux.md) - The interface should be as guessable as possible
+
+    ## Blocked By
+
+    (none)
+
+    ## Definition of Done
+
+    - [ ] Identified the application type (web, terminal, hybrid, or no UI)
+    - [ ] Listed key user scenarios
+    - [ ] Captured screenshots or output at each stage of key scenarios
+    - [ ] Reviewed evidence for UX issues
+    - [ ] Documented findings with evidence and recommendations
+    - [ ] Included verification criteria for each issue
+    - [ ] Created ideas for any UX improvements needed
+  `;
+}
 var stockAuditFunctions = {
   "agent-developer-experience": agentDeveloperExperience,
   "component-reuse": componentReuse,
@@ -1334,9 +1462,11 @@ var stockAuditFunctions = {
   "refactoring-opportunities": refactoringOpportunities,
   "repository-context": repositoryContext,
   "security-review": securityReview,
+  "slow-tests": slowTests,
   "stale-ideas": staleIdeas,
   "test-coverage": testCoverage,
-  "ubiquitous-language": ubiquitousLanguage
+  "ubiquitous-language": ubiquitousLanguage,
+  "ux-audit": uxAudit
 };
 function loadStockAudits() {
   return Object.entries(stockAuditFunctions).sort(([a], [b]) => a.localeCompare(b)).map(([name, render]) => {
@@ -2264,7 +2394,8 @@ function getRepoPath(repoName, reposDir) {
 async function cloneRepository(repository, targetPath, spawn, context) {
   return new Promise((resolve) => {
     const proc = spawn("git", ["clone", repository.gitUrl, targetPath], {
-      stdio: ["ignore", "pipe", "pipe"]
+      stdio: ["ignore", "pipe", "pipe"],
+      env: { ...process.env, GIT_TERMINAL_PROMPT: "0" }
     });
     let stderr = "";
     proc.stderr?.on("data", (data) => {
@@ -2488,26 +2619,26 @@ function getEnvironmentContext(cwd) {
 function formatLoopEvent(event) {
   switch (event.type) {
     case "loop.warning":
-      return "⚠️  WARNING: This command skips all permission checks. Only use in a sandbox environment!";
+      return "WARNING: This command skips all permission checks. Only use in a sandbox environment!";
     case "loop.started": {
       const agent2 = event.agentType ?? "claude";
-      return `\uD83D\uDD04 Starting dust loop ${agent2} (max ${event.maxIterations} iterations)...`;
+      return `Starting dust loop ${agent2} (max ${event.maxIterations} iterations)...`;
     }
     case "loop.syncing":
-      return "\uD83C\uDF0D Syncing with remote";
+      return "Syncing with remote";
     case "loop.sync_skipped":
       return `Note: git pull skipped (${event.reason})`;
     case "loop.checking_tasks":
       return null;
     case "loop.no_tasks":
-      return "\uD83D\uDE34 No tasks available. Sleeping...";
+      return "No tasks available. Sleeping...";
     case "loop.tasks_found":
-      return `✨ Found a task. Going to work!
+      return `Found a task. Going to work!
 `;
     case "loop.iteration_complete":
-      return `\uD83D\uDCCB Completed iteration ${event.iteration}/${event.maxIterations}`;
+      return `Completed iteration ${event.iteration}/${event.maxIterations}`;
     case "loop.ended":
-      return `\uD83C\uDFC1 Reached max iterations (${event.maxIterations}). Exiting.`;
+      return `Reached max iterations (${event.maxIterations}). Exiting.`;
   }
 }
 function createPostEvent(fetchFn) {
@@ -2565,7 +2696,8 @@ async function gitPull(cwd, spawn) {
   return new Promise((resolve) => {
     const proc = spawn("git", ["pull"], {
       cwd,
-      stdio: ["ignore", "pipe", "pipe"]
+      stdio: ["ignore", "pipe", "pipe"],
+      env: { ...process.env, GIT_TERMINAL_PROMPT: "0" }
     });
     let stderr = "";
     proc.stderr?.on("data", (data) => {
@@ -2815,6 +2947,9 @@ function buildEventMessage(parameters) {
     repository: parameters.repository,
     event: parameters.event
   };
+  if (parameters.repoId !== undefined) {
+    msg.repoId = parameters.repoId;
+  }
   if (parameters.agentSessionId) {
     msg.agentSessionId = parameters.agentSessionId;
   }
@@ -2900,6 +3035,7 @@ async function runRepositoryLoop(repoState, repoDeps, sendEvent, sessionId) {
         sequence,
         sessionId,
         repository: repoName,
+        repoId: repoState.repository.id,
         event,
         agentSessionId
       }));
@@ -3115,6 +3251,15 @@ function parseServerMessage(data) {
 }
 
 // lib/bucket/terminal-ui.ts
+var CHARS = {
+  dot: "*",
+  sparkle: "",
+  ellipsis: "...",
+  hline: "-",
+  arrows_lr: "<->",
+  arrows_ud: "up/dn",
+  scroll_down: "v"
+};
 var ANSI = {
   HIDE_CURSOR: "\x1B[?25l",
   SHOW_CURSOR: "\x1B[?25h",
@@ -3169,8 +3314,8 @@ function truncateLine(text, maxWidth) {
   for (let match = ansiRegex.exec(text);match !== null; match = ansiRegex.exec(text)) {
     const textBefore = text.slice(lastIndex, match.index);
     for (const char of textBefore) {
-      if (visibleCount >= maxWidth - 1) {
-        result += "…";
+      if (visibleCount >= maxWidth - CHARS.ellipsis.length) {
+        result += CHARS.ellipsis;
         return result + ANSI.RESET;
       }
       result += char;
@@ -3181,8 +3326,8 @@ function truncateLine(text, maxWidth) {
   }
   const remaining = text.slice(lastIndex);
   for (const char of remaining) {
-    if (visibleCount >= maxWidth - 1) {
-      result += "…";
+    if (visibleCount >= maxWidth - CHARS.ellipsis.length) {
+      result += CHARS.ellipsis;
       return result + ANSI.RESET;
     }
     result += char;
@@ -3282,7 +3427,7 @@ function getTabRowCount(state) {
     return 1;
   const tabWidths = [5];
   for (const name of state.repositories) {
-    tabWidths.push(name.length + 4);
+    tabWidths.push(name.length + 2 + CHARS.dot.length + 1);
   }
   let rows = 1;
   let currentRowWidth = 0;
@@ -3352,8 +3497,8 @@ function renderTabs(state) {
     const color = getRepoColor(name, i);
     const agentStatus = state.agentStatuses.get(name) ?? "idle";
     const dotColor = agentStatus === "busy" ? ANSI.FG_GREEN : ANSI.DIM;
-    const dot = `${dotColor}●${ANSI.RESET}`;
-    const width = name.length + 4;
+    const dot = `${dotColor}${CHARS.dot}${ANSI.RESET}`;
+    const width = name.length + 2 + CHARS.dot.length + 1;
     if (i === state.selectedIndex) {
       tabs.push({
         text: ` ${dot}${color} ${ANSI.INVERSE}${name}${ANSI.RESET} `,
@@ -3378,10 +3523,10 @@ function renderTabs(state) {
   return rows.map((row) => row.map((t) => t.text).join("|"));
 }
 function renderHelpLine() {
-  return `${ANSI.DIM}[←→] select  [↑↓] scroll  [PgUp/PgDn] page  [g/G] top/bottom  [o] open  [q] quit${ANSI.RESET}`;
+  return `${ANSI.DIM}[${CHARS.arrows_lr}] select  [${CHARS.arrows_ud}] scroll  [PgUp/PgDn] page  [g/G] top/bottom  [o] open  [q] quit${ANSI.RESET}`;
 }
 function renderSeparator(width) {
-  return "─".repeat(width);
+  return CHARS.hline.repeat(Math.max(0, width));
 }
 function formatLogLine(line, prefixAlign, maxWidth) {
   let prefix = "";
@@ -3413,7 +3558,7 @@ function formatLogLine(line, prefixAlign, maxWidth) {
 function renderFrame(state) {
   const lines = [];
   const hostLabel = state.connectedHost ? ` ${ANSI.DIM}[connected to ${state.connectedHost}]${ANSI.RESET}` : "";
-  lines.push(`${ANSI.BOLD}✨ dust bucket${ANSI.RESET}${hostLabel}`);
+  lines.push(`${ANSI.BOLD}${CHARS.sparkle}dust bucket${ANSI.RESET}${hostLabel}`);
   const tabRows = renderTabs(state);
   for (const tabRow of tabRows) {
     lines.push(tabRow);
@@ -3436,7 +3581,7 @@ function renderFrame(state) {
     lines.push("");
   }
   if (state.scrollOffset > 0) {
-    const indicator = `${ANSI.DIM}↓ ${state.scrollOffset} more${ANSI.RESET}`;
+    const indicator = `${ANSI.DIM}${CHARS.scroll_down} ${state.scrollOffset} more${ANSI.RESET}`;
     lines[lines.length - 1] = indicator;
   }
   const output = [];
@@ -3584,8 +3729,8 @@ function defaultSetupResize(onResize) {
 }
 function defaultGetTerminalSize() {
   return {
-    width: process.stdout.columns ?? 80,
-    height: process.stdout.rows ?? 24
+    width: process.stdout.columns || 80,
+    height: process.stdout.rows || 24
   };
 }
 function defaultWriteStdout(data) {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@joshski/dust",
-  "version": "0.1.75",
+  "version": "0.1.77",
   "description": "Flow state for AI coding agents",
   "type": "module",
   "bin": {