@tekyzinc/gsd-t 3.12.12 → 3.12.14

package/bin/gsd-t.js

@@ -1294,9 +1294,9 @@ function showInstallSummary(gsdtCount, utilCount) {
   log(`${BOLD}Quick Start:${RESET}`);
   log(`  ${DIM}$${RESET} cd your-project`);
   log(`  ${DIM}$${RESET} claude`);
-  log(`  ${DIM}>${RESET} /user:gsd-t-init my-project`);
-  log(`  ${DIM}>${RESET} /user:gsd-t-milestone "First Feature"`);
-  log(`  ${DIM}>${RESET} /user:gsd-t-wave`);
+  log(`  ${DIM}>${RESET} /gsd-t-init my-project`);
+  log(`  ${DIM}>${RESET} /gsd-t-milestone "First Feature"`);
+  log(`  ${DIM}>${RESET} /gsd-t-wave`);
   log("");
   log(`${BOLD}Other commands:${RESET}`);
   log(`  ${DIM}$${RESET} npx @tekyzinc/gsd-t status      ${DIM}— check installation${RESET}`);
@@ -1344,7 +1344,7 @@ function initClaudeMd(projectDir, projectName, today) {
         info("CLAUDE.md already contains GSD-T section — skipping");
       } else {
         warn("CLAUDE.md exists but doesn't reference GSD-T");
-        info("Run /user:gsd-t-init inside Claude Code to add GSD-T section");
+        info("Run /gsd-t-init inside Claude Code to add GSD-T section");
       }
     } else { throw e; }
   }
@@ -1513,8 +1513,8 @@ function showInitTree(projectDir) {
   log(`${BOLD}Next steps:${RESET}`);
   log(`  1. Edit CLAUDE.md — add project overview and tech stack`);
   log(`  2. Start Claude Code: ${DIM}claude${RESET}`);
-  log(`  3. Run: ${DIM}/user:gsd-t-populate${RESET}  ${DIM}(if existing codebase)${RESET}`);
-  log(`     Or:  ${DIM}/user:gsd-t-project${RESET}   ${DIM}(if new project)${RESET}`);
+  log(`  3. Run: ${DIM}/gsd-t-populate${RESET}  ${DIM}(if existing codebase)${RESET}`);
+  log(`     Or:  ${DIM}/gsd-t-project${RESET}   ${DIM}(if new project)${RESET}`);
   log("");
 }
 
@@ -1687,7 +1687,7 @@ function showStatusProject() {
     }
   } else if (hasClaudeMd) {
     info("CLAUDE.md found but no .gsd-t/ directory");
-    info("Run /user:gsd-t-init inside Claude Code to set up");
+    info("Run /gsd-t-init inside Claude Code to set up");
   } else {
     info("Not in a GSD-T project directory");
     info(`Run 'npx @tekyzinc/gsd-t init' to set up this directory`);
@@ -1945,6 +1945,19 @@ function exportUniversalRulesForNpm() {
 }
 
 async function doUpdateAll() {
+  // Step 1: Upgrade the globally-installed npm package FIRST. Without this,
+  // `update-all` would only propagate command files — but the global `gsd-t`
+  // binary itself could stay pinned to an older version (e.g., user stuck
+  // on v3.11.11 while npm registry has v3.12.12). See CHANGELOG v3.12.13.
+  // Guard with an env flag to prevent re-exec loops.
+  if (!process.env.GSDT_POST_UPGRADE) {
+    const upgraded = await upgradeGlobalBinary();
+    if (upgraded.reexec) {
+      reexecUpdateAll();
+      return;
+    }
+  }
+
   await updateGlobalCommands();
   heading("Updating registered projects...");
   log("");
@@ -1969,6 +1982,65 @@ async function doUpdateAll() {
   showUpdateAllSummary(projects.length, counts, playwrightMissing, swaggerMissing, syncCount);
 }
 
+// Upgrade the globally-installed @tekyzinc/gsd-t to @latest. Returns
+// { upgraded: bool, reexec: bool, error?: string }.
+// - reexec=true when the on-disk version after `npm install -g` is newer than
+//   the currently-running PKG_VERSION, meaning we need to hand off to the
+//   freshly-installed binary so the new code drives propagation.
+// - reexec=false when already at latest, or when the install failed (we
+//   continue with the current binary and still propagate command files).
+async function upgradeGlobalBinary() {
+  heading("Upgrading global @tekyzinc/gsd-t to latest...");
+  try {
+    execFileSync("npm", ["install", "-g", "@tekyzinc/gsd-t@latest"], {
+      stdio: "inherit",
+      env: process.env,
+    });
+  } catch (e) {
+    warn(`Global npm install failed: ${e.message || e}`);
+    info("Continuing with current binary — command files will still be propagated.");
+    return { upgraded: false, reexec: false, error: String(e.message || e) };
+  }
+
+  // Read the freshly-installed global package's version. If it's newer than
+  // the currently-running process, signal a re-exec.
+  let newVersion = null;
+  try {
+    const prefix = execFileSync("npm", ["prefix", "-g"], { encoding: "utf8" }).trim();
+    const globalPkgJson = path.join(prefix, "lib", "node_modules", "@tekyzinc", "gsd-t", "package.json");
+    if (fs.existsSync(globalPkgJson)) {
+      newVersion = JSON.parse(fs.readFileSync(globalPkgJson, "utf8")).version;
+    }
+  } catch {
+    // Best-effort; fall through.
+  }
+
+  if (newVersion && newVersion !== PKG_VERSION) {
+    success(`Global binary upgraded: v${PKG_VERSION} → v${newVersion}`);
+    info("Handing off to the newly-installed binary for propagation...");
+    return { upgraded: true, reexec: true };
+  }
+  success(`Global binary already at latest (v${PKG_VERSION})`);
+  return { upgraded: true, reexec: false };
+}
+
+// Hand execution to the newly-installed global `gsd-t update-all`. Sets the
+// GSDT_POST_UPGRADE env flag so the child does not recurse into another
+// upgrade attempt.
+function reexecUpdateAll() {
+  const env = Object.assign({}, process.env, { GSDT_POST_UPGRADE: "1" });
+  try {
+    execFileSync("gsd-t", ["update-all"], { stdio: "inherit", env });
+  } catch (e) {
+    // Surface the child's exit code; execFileSync throws with .status on
+    // non-zero exit. Fall through to re-throw so the caller exits cleanly.
+    if (e && typeof e.status === "number") {
+      process.exit(e.status);
+    }
+    throw e;
+  }
+}
+
 async function updateGlobalCommands() {
   if (getInstalledVersion() !== PKG_VERSION) {
     await doInstall({ update: true });
@@ -2671,8 +2743,10 @@ function appendHeadlessTokenLog(projectDir, entry) {
   try {
     const logPath = path.join(projectDir, ".gsd-t", "token-log.md");
     const note = entry.exitCode === 0 ? "headless exec: ok" : `headless exec: exit ${entry.exitCode}`;
+    // v3.12.14: prefer env-var model over hardcoded "unknown".
+    const model = process.env.GSD_T_MODEL || "unknown";
     const row =
-      `| ${entry.dtStart} | ${entry.dtEnd} | ${entry.command} | headless | unknown | ${entry.durationS}s | ${note} | - | - | unknown |\n`;
+      `| ${entry.dtStart} | ${entry.dtEnd} | ${entry.command} | headless | ${model} | ${entry.durationS}s | ${note} | - | - | unknown |\n`;
     const gsdtDir = path.join(projectDir, ".gsd-t");
     if (!fs.existsSync(gsdtDir)) fs.mkdirSync(gsdtDir, { recursive: true });
     if (!fs.existsSync(logPath)) {
@@ -2718,7 +2792,7 @@ function parseHeadlessFlags(args) {
  * Build the claude -p invocation string for a GSD-T command.
  *
  * Non-interactive `claude -p` mode requires the bare `/gsd-t-X` form — the
- * `/user:gsd-t-X` namespace prefix is rejected as "Unknown command" even
+ * `/gsd-t-X` namespace prefix is rejected as "Unknown command" even
  * though interactive mode accepts both. Verified by M36 Phase 0 Spike A
  * (2026-04-15). See .gsd-t/M36-spike-findings.md.
  */
@@ -2807,12 +2881,17 @@ function doHeadlessExec(command, cmdArgs, flags) {
   let output = "";
   let processExitCode = 0;
 
-  // Inject command/phase env vars so worker event-stream entries are tagged
-  // (Fix 2, v3.12.12).
+  // Inject command/phase/trace/model/project-dir env vars so worker
+  // event-stream entries (writer CLI + heartbeat hook) are tagged in the
+  // child's context (Fix 2, v3.12.12; trace/model/project-dir added v3.12.14
+  // for the null-telemetry regression fix).
   const workerEnv = Object.assign({}, process.env, {
     GSD_T_COMMAND: `gsd-t-${command}`,
     GSD_T_PHASE: process.env.GSD_T_PHASE || "execute",
+    GSD_T_PROJECT_DIR: process.env.GSD_T_PROJECT_DIR || process.cwd(),
   });
+  if (process.env.GSD_T_TRACE_ID) workerEnv.GSD_T_TRACE_ID = process.env.GSD_T_TRACE_ID;
+  if (process.env.GSD_T_MODEL) workerEnv.GSD_T_MODEL = process.env.GSD_T_MODEL;
 
   try {
     const result = execFileSync("claude", ["-p", "--dangerously-skip-permissions", prompt], {
@@ -3097,9 +3176,22 @@ function getEscalationModel(iteration) {
  */
 function spawnClaudeSession(prompt, model) {
   try {
+    // v3.12.14: propagate GSD_T_* env vars so the worker's heartbeat hook +
+    // event-writer entries are tagged with the parent command/phase/trace and
+    // this session's model. Without these, tool_call events from the debug
+    // worker appear as command=null/phase=null.
+    const env = Object.assign({}, process.env, {
+      GSD_T_COMMAND: process.env.GSD_T_COMMAND || "gsd-t-debug",
+      GSD_T_PHASE: process.env.GSD_T_PHASE || "debug",
+      GSD_T_MODEL: model || process.env.GSD_T_MODEL || null,
+      GSD_T_PROJECT_DIR: process.env.GSD_T_PROJECT_DIR || process.cwd(),
+    });
+    if (env.GSD_T_MODEL === null) delete env.GSD_T_MODEL;
+    if (process.env.GSD_T_TRACE_ID) env.GSD_T_TRACE_ID = process.env.GSD_T_TRACE_ID;
     return execFileSync("claude", ["-p", prompt, "--model", model], {
       encoding: "utf8", timeout: 300000,
       stdio: ["pipe", "pipe", "pipe"],
+      env,
     });
   } catch (e) {
     return (e.stdout || "") + (e.stderr || "") || null;
@@ -3136,8 +3228,17 @@ function runLedgerCompaction(projectDir, jsonMode) {
     JSON.stringify(entries, null, 2);
   let summary = "Compacted — see previous entries.";
   try {
+    // v3.12.14: propagate GSD_T_* env for telemetry tagging.
+    const env = Object.assign({}, process.env, {
+      GSD_T_COMMAND: process.env.GSD_T_COMMAND || "gsd-t-debug",
+      GSD_T_PHASE: process.env.GSD_T_PHASE || "debug",
+      GSD_T_MODEL: "haiku",
+      GSD_T_PROJECT_DIR: process.env.GSD_T_PROJECT_DIR || projectDir,
+    });
+    if (process.env.GSD_T_TRACE_ID) env.GSD_T_TRACE_ID = process.env.GSD_T_TRACE_ID;
     const out = execFileSync("claude", ["-p", compactPrompt, "--model", "haiku"], {
       encoding: "utf8", timeout: 120000, stdio: ["pipe", "pipe", "pipe"],
+      env,
     });
     summary = (out || "").trim() || summary;
   } catch (e) {
@@ -3378,8 +3479,8 @@ function showHelp() {
   log(`  ${DIM}$${RESET} npx @tekyzinc/gsd-t init my-saas-app`);
   log(`  ${DIM}$${RESET} npx @tekyzinc/gsd-t update\n`);
   log(`${BOLD}After installing, use in Claude Code:${RESET}`);
-  log(`  ${DIM}>${RESET} /user:gsd-t-project "Build a task management app"`);
-  log(`  ${DIM}>${RESET} /user:gsd-t-wave\n`);
+  log(`  ${DIM}>${RESET} /gsd-t-project "Build a task management app"`);
+  log(`  ${DIM}>${RESET} /gsd-t-wave\n`);
   log(`${DIM}Docs: https://github.com/Tekyz-Inc/get-stuff-done-teams${RESET}\n`);
 }
 

package/bin/headless-auto-spawn.cjs

@@ -135,12 +135,27 @@ function autoSpawnHeadless(opts) {
     const gsdtCli = path.join(projectDir, "bin", "gsd-t.js");
     const childArgs = [gsdtCli, "headless", stripGsdtPrefix(command), ...args, "--log"];
 
-    // Inject command/phase into worker env so event-stream entries are tagged
-    // (Fix 2, v3.12.12). GSD_T_PHASE defaults to "execute" for primary spawns.
+    // Inject command/phase/trace/model into worker env so event-stream entries
+    // (both the writer CLI and the heartbeat PostToolUse hook) are tagged in
+    // the child's context (Fix 2, v3.12.12; trace/model/project-dir added
+    // v3.12.14 to close the null-telemetry regression).
+    //
+    // GSD_T_PHASE defaults to "execute" for primary spawns.
+    // GSD_T_TRACE_ID / GSD_T_MODEL are inherited from parent env if set;
+    // parents (orchestrator, command files) should set them before spawning.
+    // GSD_T_PROJECT_DIR gives the writer a stable target when the child cwd
+    // drifts (e.g., temp dirs in subagents).
     const workerEnv = Object.assign({}, process.env, {
       GSD_T_COMMAND: command,
       GSD_T_PHASE: process.env.GSD_T_PHASE || "execute",
+      GSD_T_PROJECT_DIR: process.env.GSD_T_PROJECT_DIR || projectDir,
     });
+    if (process.env.GSD_T_TRACE_ID) {
+      workerEnv.GSD_T_TRACE_ID = process.env.GSD_T_TRACE_ID;
+    }
+    if (process.env.GSD_T_MODEL) {
+      workerEnv.GSD_T_MODEL = process.env.GSD_T_MODEL;
+    }
 
     const child = spawn("node", childArgs, {
       cwd: projectDir,
@@ -392,8 +407,12 @@ function appendTokenLog(projectDir, entry) {
   try {
     const logPath = path.join(projectDir, ".gsd-t", "token-log.md");
     const note = entry.exitCode === 0 ? "headless spawn: ok" : `headless spawn: exit ${entry.exitCode}`;
+    // v3.12.14: prefer env-var model over the old hardcoded "unknown". The
+    // parent session should have set GSD_T_MODEL before invoking spawn; fall
+    // back to "unknown" if not (graceful).
+    const model = process.env.GSD_T_MODEL || "unknown";
     const row =
-      `| ${entry.dtStart} | ${entry.dtEnd} | ${entry.command} | headless | unknown | ${entry.durationS}s | ${note} | - | - | unknown |\n`;
+      `| ${entry.dtStart} | ${entry.dtEnd} | ${entry.command} | headless | ${model} | ${entry.durationS}s | ${note} | - | - | unknown |\n`;
     if (!fs.existsSync(logPath)) {
       // Create with header
       ensureDir(path.dirname(logPath));

package/bin/headless-auto-spawn.js

@@ -135,11 +135,22 @@ function autoSpawnHeadless(opts) {
     const gsdtCli = path.join(projectDir, "bin", "gsd-t.js");
     const childArgs = [gsdtCli, "headless", stripGsdtPrefix(command), ...args, "--log"];
 
+    // v3.12.14 — env-var propagation for telemetry tagging (mirror of the
+    // authoritative .cjs variant). PRODUCTION consumers import the .cjs; this
+    // .js copy is retained only to keep the legacy test green.
+    const workerEnv = Object.assign({}, process.env, {
+      GSD_T_COMMAND: command,
+      GSD_T_PHASE: process.env.GSD_T_PHASE || "execute",
+      GSD_T_PROJECT_DIR: process.env.GSD_T_PROJECT_DIR || projectDir,
+    });
+    if (process.env.GSD_T_TRACE_ID) workerEnv.GSD_T_TRACE_ID = process.env.GSD_T_TRACE_ID;
+    if (process.env.GSD_T_MODEL) workerEnv.GSD_T_MODEL = process.env.GSD_T_MODEL;
+
     const child = spawn("node", childArgs, {
       cwd: projectDir,
       detached: true,
       stdio: ["ignore", logFd, logFd],
-      env: process.env,
+      env: workerEnv,
     });
 
     child.unref();

package/bin/orchestrator.js

@@ -82,6 +82,25 @@ function isPortInUse(port) {
   }
 }
 
+/**
+ * Build the env block for orchestrator-spawned `claude -p` sessions.
+ * v3.12.14 — propagates GSD_T_COMMAND/PHASE/TRACE_ID/MODEL/PROJECT_DIR so the
+ * worker's heartbeat hook and event-writer entries are tagged. Accepts the
+ * opts object (for `label`) that callers already pass and defaults missing
+ * fields from the parent process env.
+ */
+function _buildOrchestratorEnv(opts, projectDir) {
+  opts = opts || {};
+  const env = Object.assign({}, process.env, {
+    GSD_T_COMMAND: process.env.GSD_T_COMMAND || "gsd-t-orchestrator",
+    GSD_T_PHASE: process.env.GSD_T_PHASE || opts.label || "orchestrator",
+    GSD_T_PROJECT_DIR: process.env.GSD_T_PROJECT_DIR || projectDir,
+  });
+  if (process.env.GSD_T_TRACE_ID) env.GSD_T_TRACE_ID = process.env.GSD_T_TRACE_ID;
+  if (process.env.GSD_T_MODEL) env.GSD_T_MODEL = process.env.GSD_T_MODEL;
+  return env;
+}
+
 // ─── Orchestrator Class ────────────────────────────────────────────────────
 
 class Orchestrator {
@@ -212,11 +231,15 @@ ${BOLD}Phases:${RESET} ${this.wf.phases.join(" → ")}
     const signalFile = path.join(this.getReviewDir(projectDir), `_sync-done-${Date.now()}.json`);
     let result = { output: "", exitCode: 1, duration: 0 };
 
+    // v3.12.14: propagate GSD_T_* env so the worker's heartbeat hook and
+    // event-writer entries are tagged. opts.label → phase/command hints.
+    const env = _buildOrchestratorEnv(opts, projectDir);
     const child = execFile("claude", args, {
       encoding: "utf8",
       timeout: effectiveTimeout,
       cwd: projectDir,
       maxBuffer: 10 * 1024 * 1024,
+      env,
     }, (err, stdout, stderr) => {
       this.untrackChild(child.pid);
       const raw = err ? ((err.stdout || "") + (err.stderr || "")) : (stdout || "");
@@ -264,12 +287,15 @@ ${BOLD}Phases:${RESET} ${this.wf.phases.join(" → ")}
       );
     }
 
+    // v3.12.14: propagate GSD_T_* env for telemetry tagging in the child.
+    const env = _buildOrchestratorEnv(opts, projectDir);
     return new Promise((resolve) => {
       const child = execFile("claude", args, {
         encoding: "utf8",
         timeout: timeout || 120_000,
         cwd: projectDir,
         maxBuffer: 10 * 1024 * 1024,
+        env,
       }, (err, stdout, stderr) => {
         this.untrackChild(child.pid);
         const raw = err ? ((err.stdout || "") + (err.stderr || "")) : (stdout || "");

package/commands/gsd-t-backlog-promote.md

@@ -13,14 +13,14 @@ Parse $ARGUMENTS to extract:
 - `<position>` — the entry number to promote
 
 If no position is provided, show an error:
-"Usage: `/user:gsd-t-backlog-promote <position>`"
+"Usage: `/gsd-t-backlog-promote <position>`"
 
 ## Step 2: Find and Display Entry
 
 Find the entry at the specified position in `.gsd-t/backlog.md`.
 
 If the position doesn't exist, show an error:
-"No backlog entry at position {position}. Run `/user:gsd-t-backlog-list` to see available entries."
+"No backlog entry at position {position}. Run `/gsd-t-backlog-list` to see available entries."
 
 Display the entry to the user:
 ```
@@ -88,10 +88,10 @@ After classification is confirmed:
 
 Based on the classification, present the command for the user to invoke:
 
-- **Milestone**: "Run `/user:gsd-t-milestone {refined description}`"
-- **Quick**: "Run `/user:gsd-t-quick {refined description}`"
-- **Debug**: "Run `/user:gsd-t-debug {refined description}`"
-- **Feature**: "Run `/user:gsd-t-feature {refined description}`"
+- **Milestone**: "Run `/gsd-t-milestone {refined description}`"
+- **Quick**: "Run `/gsd-t-quick {refined description}`"
+- **Debug**: "Run `/gsd-t-debug {refined description}`"
+- **Feature**: "Run `/gsd-t-feature {refined description}`"
 
 Display the full command with the refined description ready to copy.
 

package/commands/gsd-t-complete-milestone.md

@@ -3,8 +3,8 @@
 You are finalizing a completed milestone. Your job is to archive the milestone documentation, create a git tag, and prepare for the next milestone.
 
 This command is:
-- **Auto-invoked** by `/user:gsd-t-verify` (Step 8) after all quality gates pass — at ALL autonomy levels
-- **Auto-invoked** by `/user:gsd-t-wave` as part of the VERIFY+COMPLETE phase
+- **Auto-invoked** by `/gsd-t-verify` (Step 8) after all quality gates pass — at ALL autonomy levels
+- **Auto-invoked** by `/gsd-t-wave` as part of the VERIFY+COMPLETE phase
 - **Standalone** when user wants to manually close a milestone
 
 ## Step 1: Verify Completion
@@ -14,7 +14,7 @@ Read:
 2. `.gsd-t/verify-report.md` — confirm all checks passed
 
 If status is not VERIFIED:
-"⚠️ Milestone not yet verified. Run `/user:gsd-t-verify` first, or use `--force` to complete anyway."
+"⚠️ Milestone not yet verified. Run `/gsd-t-verify` first, or use `--force` to complete anyway."
 
 If `--force` flag provided, proceed with warning in archive.
 
@@ -76,7 +76,7 @@ Refer to `.gsd-t/contracts/goal-backward-contract.md` for the full verification
   {findings table}
 
   Options:
-  1. Fix the findings and re-run /user:gsd-t-verify
+  1. Fix the findings and re-run /gsd-t-verify
   2. Override with explicit acknowledgment: re-run this command with --force-goal-backward
 
   Proceed with option 1 (recommended) or acknowledge to force completion?
@@ -493,8 +493,8 @@ Summary:
 
 Next steps:
 - Push tags: git push origin v{version}
-- Start next milestone: /user:gsd-t-milestone "{next name}"
-- Or view roadmap: /user:gsd-t-status
+- Start next milestone: /gsd-t-milestone "{next name}"
+- Or view roadmap: /gsd-t-status
 ```
 
 ## Step 13: Update Roadmap (if exists)
@@ -510,7 +510,7 @@ If `.gsd-t/roadmap.md` exists:
 "Cannot complete — verification found issues. Address them first or use `--force`."
 
 ### If no milestone active:
-"No active milestone to complete. Run `/user:gsd-t-status` to see state."
+"No active milestone to complete. Run `/gsd-t-status` to see state."
 
 ### If git operations fail:
 - Still create archive

package/commands/gsd-t-design-audit.md

@@ -400,7 +400,7 @@ If ANY CRITICAL or HIGH deviations were found, automatically prompt the fix work
 The audit report at `.gsd-t/design-audit-{page-name}-{YYYY-MM-DD}.md`
 has the exact Figma values for each deviation.
 
-`/user:gsd-t-quick fix all CRITICAL and HIGH deviations from .gsd-t/design-audit-{page-name}-{YYYY-MM-DD}.md — use the Figma values in the report as the source of truth`
+`/gsd-t-quick fix all CRITICAL and HIGH deviations from .gsd-t/design-audit-{page-name}-{YYYY-MM-DD}.md — use the Figma values in the report as the source of truth`
 
 ───────────────────────────────────────────────────────────────
 ```
@@ -414,7 +414,7 @@ If ONLY MEDIUM or LOW deviations remain, show:
 
 **{N} MEDIUM + {N} LOW deviations.** These are minor — fix if you want pixel-perfect.
 
-`/user:gsd-t-quick fix MEDIUM and LOW deviations from .gsd-t/design-audit-{page-name}-{YYYY-MM-DD}.md`
+`/gsd-t-quick fix MEDIUM and LOW deviations from .gsd-t/design-audit-{page-name}-{YYYY-MM-DD}.md`
 
 ───────────────────────────────────────────────────────────────
 ```
@@ -427,7 +427,7 @@ After fixes are applied, **re-run the audit automatically** to verify. Loop unti
 
 ## Rules
 
-- **You write ZERO code during the audit phase (Steps 1-5).** Report only. Code changes happen in Step 6 via `/user:gsd-t-quick`.
+- **You write ZERO code during the audit phase (Steps 1-5).** Report only. Code changes happen in Step 6 via `/gsd-t-quick`.
 - **You do NOT "look close" at anything.** Every property gets an exact value from Figma and an exact value from the build. They match or they don't.
 - **You do NOT skip widgets.** Every widget in the Figma AND every widget in the build gets audited.
 - **You MUST call `get_design_context` per widget node — NOT `get_screenshot`.** `get_design_context` returns structured code, component properties, and design tokens. `get_screenshot` returns only a visual image that you cannot extract exact values from. Using `get_screenshot` for widget extraction defeats the entire purpose of structured comparison — you end up eyeballing instead of measuring. The ONLY acceptable use of `get_screenshot` is for the built page (Step 2) where you need to see what was actually rendered. For Figma source data, ALWAYS use `get_design_context`.

package/commands/gsd-t-design-build.md

@@ -40,7 +40,7 @@ Pass any of these as `$ARGUMENTS`:
 ## Prerequisites
 
 - Design contracts must exist in `.gsd-t/contracts/design/` with an `INDEX.md`
-- If no contracts exist, run `/user:gsd-t-design-decompose` first
+- If no contracts exist, run `/gsd-t-design-decompose` first
 
 ## Why a JS Orchestrator?
 

package/commands/gsd-t-design-decompose.md

@@ -471,7 +471,7 @@ This domain owns the following design contracts:
 - Pages: (none — pages owned by page-assembly domain)
 ```
 
-If `.gsd-t/domains/` does NOT exist yet, suggest the user run `/user:gsd-t-partition` next, with a note that design contracts should be partitioned into domains:
+If `.gsd-t/domains/` does NOT exist yet, suggest the user run `/gsd-t-partition` next, with a note that design contracts should be partitioned into domains:
 - **design-system domain** owns element contracts
 - **widgets domain** owns widget contracts
 - **pages domain** owns page assembly + routing
@@ -494,11 +494,11 @@ Display:
 
 **Design Build** — build UI from contracts with tiered review gates (elements → widgets → pages)
 
-`/user:gsd-t-design-build`
+`/gsd-t-design-build`
 
 **Also available:**
-- `/user:gsd-t-partition` — if you need domain boundaries before building
-- `/user:gsd-t-plan` — if you need task lists before building
+- `/gsd-t-partition` — if you need domain boundaries before building
+- `/gsd-t-plan` — if you need task lists before building
 
 ───────────────────────────────────────────────────────────────
 ```

package/commands/gsd-t-execute.md

@@ -66,7 +66,7 @@ This calls `getSessionStatus()` (v2.0.0) which reads `.gsd-t/.context-meter-stat
 
 Use the returned `threshold` as the gate signal for the rest of this run. The gate logic is in Step 3.5; do NOT skip it. If the Context Meter hook isn't installed (`.gsd-t/.context-meter-state.json` missing and doctor reports it), run `gsd-t doctor` to diagnose — the gate still works via the heuristic fallback but real-time readings give much better guardrails.
 
-Why: every `/user:gsd-t-execute` invocation is a fresh orchestrator session and needs a current reading of context utilization before spawning any subagents. The authoritative source is the Context Meter state file; the fallback keeps the gate functional on projects that haven't installed the hook yet.
+Why: every `/gsd-t-execute` invocation is a fresh orchestrator session and needs a current reading of context utilization before spawning any subagents. The authoritative source is the Context Meter state file; the fallback keeps the gate functional on projects that haven't installed the hook yet.
 
 ## Step 1: Load State
 

package/commands/gsd-t-feature.md

@@ -1,6 +1,6 @@
 # GSD-T: Feature — Add a Major Feature to an Existing Project
 
-You are the lead agent planning a significant new feature for an existing codebase. Unlike `/user:gsd-t-project` (greenfield), this command respects and builds on what already exists — existing patterns, schema, auth, conventions, and contracts.
+You are the lead agent planning a significant new feature for an existing codebase. Unlike `/gsd-t-project` (greenfield), this command respects and builds on what already exists — existing patterns, schema, auth, conventions, and contracts.
 
 ## Step 0.5: Scan Freshness Auto-Refresh
 
@@ -154,7 +154,7 @@ The feature may be a single milestone or multiple, depending on scope:
 ### Single Milestone (if feature is focused):
 - 2-4 domains, < 15 tasks total
 - Minimal impact on existing contracts
-- Skip roadmap, go straight to: "Run `/user:gsd-t-partition` to begin"
+- Skip roadmap, go straight to: "Run `/gsd-t-partition` to begin"
 
 ### Multiple Milestones (if feature is complex):
 Apply these sequencing rules:
@@ -267,7 +267,7 @@ Present:
 4. Any conflicts with existing work
 5. Recommended starting point
 
-Ask: "Ready to start? Run `/user:gsd-t-partition` for Milestone {N}."
+Ask: "Ready to start? Run `/gsd-t-partition` for Milestone {N}."
 
 $ARGUMENTS
 

package/commands/gsd-t-gap-analysis.md

@@ -224,13 +224,13 @@ Show the recommended groupings and offer promotion paths:
 ## Recommended Next Steps
 
 1. {Milestone name} — {N} gaps ({critical} critical, {high} high)
-   → /user:gsd-t-milestone "{name}"
+   → /gsd-t-milestone "{name}"
 
 2. {Feature name} — {N} gaps
-   → /user:gsd-t-feature "{name}"
+   → /gsd-t-feature "{name}"
 
 3. Quick fixes — {N} items
-   → /user:gsd-t-quick "{description}"
+   → /gsd-t-quick "{description}"
 
 Promote any of these now, or review the gap analysis first?
 ```

package/commands/gsd-t-health.md

@@ -131,9 +131,9 @@ After repair, re-run the checks and report the final state.
 ## Step 6: Next Steps
 
 If HEALTHY → "✅ GSD-T structure is healthy — all required files present."
-If DEGRADED with --repair done → "✅ Repaired {N} missing files. Run /user:gsd-t-health again to confirm."
-If DEGRADED without --repair → "⚠ Run /user:gsd-t-health --repair to create {N} missing files."
-If BROKEN → "🔴 Project structure is broken. Run /user:gsd-t-health --repair or /user:gsd-t-init to rebuild."
+If DEGRADED with --repair done → "✅ Repaired {N} missing files. Run /gsd-t-health again to confirm."
+If DEGRADED without --repair → "⚠ Run /gsd-t-health --repair to create {N} missing files."
+If BROKEN → "🔴 Project structure is broken. Run /gsd-t-health --repair or /gsd-t-init to rebuild."
 
 $ARGUMENTS
 

package/commands/gsd-t-help.md

@@ -84,8 +84,8 @@ BACKLOG                                                                Manual
   backlog-settings    Manage types, apps, categories, and defaults
 
 ───────────────────────────────────────────────────────────────────────────────
-Type /user:gsd-t-help {command} for detailed help on any command.
-Example: /user:gsd-t-help impact
+Type /gsd-t-help {command} for detailed help on any command.
+Example: /gsd-t-help impact
 ╚══════════════════════════════════════════════════════════════════════════════╝
 ```
 
@@ -94,7 +94,7 @@ Example: /user:gsd-t-help impact
 When user provides a command name, show detailed help:
 
 ```
-/user:gsd-t-help {command}
+/gsd-t-help {command}
 ```
 
 ### Command Details Format:
@@ -108,7 +108,7 @@ SUMMARY
   {One-line description}
 
 USAGE
-  /user:gsd-t-{command} [arguments]
+  /gsd-t-{command} [arguments]
 
 WHEN TO USE
   • {Use case 1}
@@ -131,8 +131,8 @@ FILES CREATED/UPDATED
   • {file 2}
 
 EXAMPLES
-  /user:gsd-t-{command}
-  /user:gsd-t-{command} "with argument"
+  /gsd-t-{command}
+  /gsd-t-{command} "with argument"
 
 RELATED COMMANDS
   • {related 1} — {why related}
@@ -149,10 +149,10 @@ Use these when user asks for help on a specific command:
 - **Summary**: Describe what you need in plain language — auto-routes to the right GSD-T command using semantic evaluation
 - **Auto-invoked**: Yes — via UserPromptSubmit hook when prompt does not start with `/`
 - **Files**: Reads `CLAUDE.md`, `.gsd-t/progress.md`, command summaries from `gsd-t-help`
-- **How it works**: First checks if this is a continuation of an ongoing command (mid-task follow-up, status report, or acknowledgment) — if so, outputs `→ /gsd ──▶ continue /user:gsd-t-{last-command}` and resumes. For new requests, evaluates against every command's purpose and "Use when" criteria. Commands that match get shortlisted, best fit is selected. Shows runner-up when close.
+- **How it works**: First checks if this is a continuation of an ongoing command (mid-task follow-up, status report, or acknowledgment) — if so, outputs `→ /gsd ──▶ continue /gsd-t-{last-command}` and resumes. For new requests, evaluates against every command's purpose and "Use when" criteria. Commands that match get shortlisted, best fit is selected. Shows runner-up when close.
 - **Auto-route**: After `gsd-t install`, any plain text message (no leading `/`) is automatically routed through `/gsd`. Slash commands pass through unchanged. Binary detection — no heuristics.
 - **Use when**: You don't want to remember which command to use — just describe what you want
-- **Examples**: `/user:gsd Fix the login bug`, `/user:gsd Add dark mode`, `/user:gsd Scan for tech debt`
+- **Examples**: `/gsd Fix the login bug`, `/gsd Add dark mode`, `/gsd Scan for tech debt`
 - **Auto-route examples**: `Fix the login bug` (no slash needed), `Add dark mode`, `Scan for tech debt`
 
 ### prd
@@ -228,7 +228,7 @@ Use these when user asks for help on a specific command:
 - **Use when**: Ready to define specific implementation tasks
 - **Note (M22)**: Tasks auto-split if estimated scope exceeds 70% context window — guarantees fresh dispatch works
 - **Note (M26)**: Pre-mortem step now also reads rules.jsonl for historical failure patterns via getPreMortemRules
-- **Note (M38)**: Conversational use cases (formerly `/user:gsd-t-prompt`, `/user:gsd-t-brainstorm`, `/user:gsd-t-discuss`) are now handled by the Smart Router's conversational mode — just describe what you want via `/user:gsd` or plain text.
+- **Note (M38)**: Conversational use cases (formerly `/gsd-t-prompt`, `/gsd-t-brainstorm`, `/gsd-t-discuss`) are now handled by the Smart Router's conversational mode — just describe what you want via `/gsd` or plain text.
 
 ### impact
 - **Summary**: Analyze downstream effects of planned changes
@@ -433,7 +433,7 @@ Did you mean one of these?
   • {closest match 1}
   • {closest match 2}
 
-Run /user:gsd-t-help for full command list.
+Run /gsd-t-help for full command list.
 ```
 
 $ARGUMENTS