lsd-pi 1.1.2 → 1.1.3

package/src/resources/skills/{create-gsd-extension → create-lsd-extension}/workflows/debug-extension.md

@@ -24,10 +24,10 @@
 
 ```bash
 # Test in isolation
-gsd -e ./path/to/extension.ts
+lsd -e ./path/to/extension.ts
 
-# Check GSD startup output for errors
-# Extension errors are logged but don't crash GSD
+# Check LSD startup output for errors
+# Extension errors are logged but don't crash LSD
 ```
 
 ## Step 3: Verify File Location
@@ -38,7 +38,7 @@ Community extensions must be in auto-discovery paths:
 - `.gsd/extensions/*.ts`
 - `.gsd/extensions/*/index.ts`
 
-Note: `~/.gsd/agent/extensions/` is reserved for bundled extensions synced from the gsd-pi package.
+Note: `~/.gsd/agent/extensions/` is reserved for bundled extensions synced from the lsd-pi package.
 
 The file must `export default function(pi: ExtensionAPI) { ... }`.
 
@@ -49,7 +49,7 @@ Read `../references/key-rules-gotchas.md` and verify each rule against the exten
 ## Step 5: Add Debugging
 
 ```typescript
-// Temporary: log to stderr (visible in GSD output)
+// Temporary: log to stderr (visible in LSD output)
 console.error("[my-ext] Loading...");
 
 pi.on("session_start", async (_event, ctx) => {

package/src/resources/skills/teams-debug/SKILL.md

@@ -1,18 +1,17 @@
 ---
 name: teams-debug
 description: "Fix any bug related to an active Teams plan. Can be triggered by the user at any time or automatically from teams-verify on failure. Reads the plan for context, spawns a targeted builder subagent to fix the issue, then appends a debug status update to the plan."
-user-invocable: true
 ---
 
 # Teams: Debug
 
 Fix a bug related to the current Teams plan. This skill can be triggered:
-- **By the user at any time** — e.g. "something's broken, use `/teams-debug`"
-- **Automatically from `/teams-verify`** — when a scenario fails during manual verification
+- **By the user at any time** — e.g. "something's broken, use `/skill:teams-debug`"
+- **Automatically from `/skill:teams-verify`** — when a scenario fails during manual verification
 
 It reads the plan (which contains all review and verification history) before fixing.
 
-**Prerequisite:** A `.lsd/plan/PLAN-*.md` file must exist. If none found, stop and tell the user to run `/teams-plan` first.
+**Prerequisite:** A `.lsd/plan/PLAN-*.md` file must exist. If none found, stop and tell the user to run `/skill:teams-plan` first.
 
 ---
 
@@ -32,7 +31,7 @@ Otherwise, ask:
 
 > **What's the bug?**
 >
-> Describe what went wrong, what you expected, and what you saw. Include the scenario name if it came from `/teams-verify`.
+> Describe what went wrong, what you expected, and what you saw. Include the scenario name if it came from `/skill:teams-verify`.
 
 Wait for their response.
 
@@ -127,4 +126,4 @@ Print:
 ```
 
 Then ask:
-> **Want to re-verify this scenario? Run `/teams-verify` to continue from where you left off.**
+> **Want to re-verify this scenario? Run `/skill:teams-verify` to continue from where you left off.**

package/src/resources/skills/teams-document/SKILL.md

@@ -1,7 +1,6 @@
 ---
 name: teams-document
 description: "Update existing project docs (README.md, ARCHITECTURE.md, etc.) and the active plan file to reflect the latest build. Spawns a worker subagent scribe to find and update relevant documentation files and ensure the plan is accurate, then appends a docs status update to the plan."
-user-invocable: true
 ---
 
 # Teams: Document
@@ -9,7 +8,7 @@ user-invocable: true
 Update existing project documentation to reflect what was built in the current plan. Spawns a worker subagent that finds relevant docs and updates them — no new files created unless explicitly needed.
 
 **Prerequisite:** A `.lsd/plan/PLAN-*.md` file must exist. If none found, stop:
-> No plan files found in `.lsd/plan/`. Run `/teams-plan` first.
+> No plan files found in `.lsd/plan/`. Run `/skill:teams-plan` first.
 
 ---
 

package/src/resources/skills/teams-plan/SKILL.md

@@ -1,7 +1,6 @@
 ---
 name: teams-plan
 description: "Plan and build a feature. Orchestrator plans, spawns sequential builder subagents per phase (with Playwright/Maestro verification), then a reviewer subagent, then a builder to apply fixes. Supports parallel mode for independent phases."
-user-invocable: true
 ---
 
 # Teams: Plan + Build
@@ -41,7 +40,7 @@ mkdir -p .lsd/plan
 - Check for existing plan files: `.lsd/plan/PLAN-*.md`
 - Count existing files to determine N (next plan number = count + 1)
 - If plans exist, inform the user:
-  > **Found [N-1] existing plan(s). Creating Plan #[N]. Use `/teams-run` to resume an existing plan.**
+  > **Found [N-1] existing plan(s). Creating Plan #[N]. Use `/skill:teams-run` to resume an existing plan.**
 - If no existing plans: plan number = 1
 
 Write `.lsd/plan/PLAN-[N].md`:
@@ -291,7 +290,7 @@ Final summary format:
 ```
 
 Then suggest:
-> **Build done. Run `/teams-verify` to walk through manual E2E verification.**
+> **Build done. Run `/skill:teams-verify` to walk through manual E2E verification.**
 
 ---
 
@@ -299,7 +298,7 @@ Then suggest:
 
 Ask the user:
 
-> **"Would you like to update your documentation? Run `/teams-document` to have the scribe update your README, ARCHITECTURE.md, and other docs."**
+> **"Would you like to update your documentation? Run `/skill:teams-document` to have the scribe update your README, ARCHITECTURE.md, and other docs."**
 
 If yes, invoke the `teams-document` skill.
 

package/src/resources/skills/teams-run/SKILL.md

@@ -1,7 +1,6 @@
 ---
 name: teams-run
 description: "Resume building a single Teams plan. Orchestrator spawns sequential or parallel builder subagents per incomplete phase, then a reviewer subagent, then applies fixes."
-user-invocable: true
 ---
 
 # Teams: Run (Resume Build)
@@ -13,7 +12,7 @@ You are the orchestrator. Resume an existing build by running all incomplete pha
 ## Step 1: Find the Plan
 
 List all files matching `.lsd/plan/PLAN-*.md`. If none exist:
-> No plan files found in `.lsd/plan/`. Use `/teams-plan` to create a plan first.
+> No plan files found in `.lsd/plan/`. Use `/skill:teams-plan` to create a plan first.
 
 If multiple plan files exist, show the list and ask the user which plan to resume. Default to the highest-numbered plan.
 
@@ -216,7 +215,7 @@ Final summary:
 ```
 
 Then suggest:
-> **Build done. Run `/teams-verify` to walk through manual E2E verification.**
+> **Build done. Run `/skill:teams-verify` to walk through manual E2E verification.**
 
 ---
 
@@ -224,7 +223,7 @@ Then suggest:
 
 Ask the user:
 
-> **"Would you like to update your documentation? Run `/teams-document` to have the scribe update your README, ARCHITECTURE.md, and other docs."**
+> **"Would you like to update your documentation? Run `/skill:teams-document` to have the scribe update your README, ARCHITECTURE.md, and other docs."**
 
 If yes, invoke the `teams-document` skill.
 

package/src/resources/skills/teams-verify/SKILL.md

@@ -1,7 +1,6 @@
 ---
 name: teams-verify
 description: "Guide the user through manually verifying the build E2E — walks through each scenario step-by-step, records pass/fail, and appends a verification status update to the plan file."
-user-invocable: true
 ---
 
 # Teams: Manual Verify
@@ -13,7 +12,7 @@ You guide the user through manually verifying the completed build end-to-end. On
 ## Step 1: Load the Plan
 
 Find and read the most recent `.lsd/plan/PLAN-*.md` (highest number). Store as `PLAN_FILE`. If none exist:
-> No plan files found in `.lsd/plan/`. Run `/teams-plan` first.
+> No plan files found in `.lsd/plan/`. Run `/skill:teams-plan` first.
 
 Extract:
 - Plan ID (the `Plan ID:` field — e.g. `#2`)
@@ -63,7 +62,7 @@ Then ask:
 
 - **pass** → record it, move to next scenario
 - **fail** → ask: *"What went wrong?"* Record their description. Then immediately offer:
-  > **Bug detected. Run `/teams-debug` to fix it now, or continue verifying the rest first?**
+  > **Bug detected. Run `/skill:teams-debug` to fix it now, or continue verifying the rest first?**
 
   If they say **fix now**: invoke the `teams-debug` skill directly, passing the scenario name and their failure description as context. After the fix is applied, re-run this scenario before continuing.
 
@@ -115,9 +114,9 @@ Print the final result:
 If there are failures, offer:
 
 > **N scenario(s) failed. Options:**
-> - `/teams-debug` — fix bugs one at a time with full plan context (recommended)
+> - `/skill:teams-debug` — fix bugs one at a time with full plan context (recommended)
 > - **Fix all** — I spawn a single builder to address all failures at once
 
-If they choose **`/teams-debug`**: invoke the `teams-debug` skill for each failed scenario in order, passing the scenario name and failure description. After each fix, update the `## Verification` section in the plan to mark the scenario as `FIXED`.
+If they choose **`/skill:teams-debug`**: invoke the `teams-debug` skill for each failed scenario in order, passing the scenario name and failure description. After each fix, update the `## Verification` section in the plan to mark the scenario as `FIXED`.
 
 If they choose **fix all**: read the failures from the plan's `## Verification` section and spawn a **Sonnet** builder subagent with the full list of issues. Then offer to re-run verification on just the failed scenarios.

package/dist/resources/extensions/slash-commands/create-extension.js

@@ -1,264 +0,0 @@
-import { showInterviewRound } from "../shared/tui.js";
-export default function createExtension(pi) {
-    pi.registerCommand("create-extension", {
-        description: "Scaffold a new pi extension with interview-driven context gathering",
-        async handler(args, ctx) {
-            const inlineName = (typeof args === "string" ? args : "").trim();
-            // ── Interview — always runs first ─────────────────────────────────────
-            const questions = [
-                ...(!inlineName
-                    ? [
-                        {
-                            id: "purpose",
-                            header: "Purpose",
-                            question: "What should this extension do?",
-                            options: [
-                                {
-                                    label: "Add a custom tool",
-                                    description: "Register a new tool the LLM can call (like gsd_plan, plan_clarify).",
-                                },
-                                {
-                                    label: "Add a slash command",
-                                    description: "A /command the user types — runs logic, optionally triggers an agent turn.",
-                                },
-                                {
-                                    label: "React to agent events",
-                                    description: "Hook into turn_end, agent_end, tool_call, etc. to observe or intercept.",
-                                },
-                                {
-                                    label: "Custom TUI component",
-                                    description: "Render a widget, overlay, dialog, or custom editor in the terminal UI.",
-                                },
-                            ],
-                        },
-                    ]
-                    : []),
-                {
-                    id: "ui",
-                    header: "UI",
-                    question: "Does this extension need any custom UI?",
-                    options: [
-                        {
-                            label: "No UI",
-                            description: "Pure logic — no dialogs, widgets, or custom rendering needed.",
-                        },
-                        {
-                            label: "Dialogs only",
-                            description: "Uses built-in ctx.ui.select / ctx.ui.input / ctx.ui.confirm dialogs.",
-                        },
-                        {
-                            label: "Status / widget",
-                            description: "Shows a persistent status indicator or footer widget.",
-                        },
-                        {
-                            label: "Full custom component",
-                            description: "Uses ctx.ui.custom() to render a fully bespoke TUI component.",
-                        },
-                    ],
-                },
-                {
-                    id: "events",
-                    header: "Events",
-                    question: "Does it need to hook into the agent lifecycle?",
-                    options: [
-                        {
-                            label: "No — standalone",
-                            description: "Runs only when explicitly invoked — no event listeners needed.",
-                        },
-                        {
-                            label: "Yes — tool_call",
-                            description: "Intercepts or observes tool calls before or after they run.",
-                        },
-                        {
-                            label: "Yes — turn / session",
-                            description: "Reacts to turn_end, agent_end, session_start, or similar lifecycle events.",
-                        },
-                        {
-                            label: "Yes — context / prompt",
-                            description: "Modifies the system prompt or filters messages via context / before_agent_start.",
-                        },
-                    ],
-                },
-                {
-                    id: "persistence",
-                    header: "State",
-                    question: "Does this extension need to persist state across sessions?",
-                    options: [
-                        {
-                            label: "No state needed",
-                            description: "Stateless — each invocation is independent.",
-                        },
-                        {
-                            label: "In-memory only",
-                            description: "Keeps state while the session is running but doesn't survive restarts.",
-                        },
-                        {
-                            label: "Persisted to session",
-                            description: "Uses pi.appendEntry() to write state into the session JSONL for resume.",
-                        },
-                    ],
-                },
-                {
-                    id: "complexity",
-                    header: "Complexity",
-                    question: "How complex is the implementation?",
-                    options: [
-                        {
-                            label: "Simple — one concern",
-                            description: "A single tool or command, minimal branching, easy to follow.",
-                        },
-                        {
-                            label: "Moderate — a few parts",
-                            description: "A command plus an event hook, or a tool with custom rendering.",
-                        },
-                        {
-                            label: "Complex — full extension",
-                            description: "Multiple tools, commands, events, UI, and state working together.",
-                        },
-                    ],
-                },
-            ];
-            const result = await showInterviewRound(questions, {
-                progress: "New pi extension · Context",
-                reviewHeadline: "Review your choices",
-                exitHeadline: "Cancel extension creation?",
-                exitLabel: "cancel",
-            }, ctx);
-            // User hit Esc — bail silently
-            if (!result.answers || Object.keys(result.answers).length === 0) {
-                ctx.ui.notify("Cancelled.", "info");
-                return;
-            }
-            // ── Resolve name / description ────────────────────────────────────────
-            let extensionDescription = inlineName;
-            if (!extensionDescription) {
-                const purpose = result.answers["purpose"];
-                if (purpose) {
-                    extensionDescription = purpose.notes?.trim()
-                        ? purpose.notes.trim()
-                        : Array.isArray(purpose.selected) ? purpose.selected[0] : purpose.selected;
-                }
-            }
-            if (!extensionDescription) {
-                ctx.ui.notify("No description captured — add details in the notes field next time.", "warning");
-                return;
-            }
-            // ── Build and send the enriched prompt ────────────────────────────────
-            sendPrompt(extensionDescription, result, pi);
-        },
-    });
-}
-// ─── Prompt builder ───────────────────────────────────────────────────────────
-function formatAnswers(result) {
-    const lines = [];
-    const purpose = result.answers["purpose"];
-    if (purpose?.notes) {
-        lines.push(`- **Extension goal (user's words)**: ${purpose.notes}`);
-    }
-    const ui = result.answers["ui"];
-    if (ui) {
-        const selected = Array.isArray(ui.selected) ? ui.selected[0] : ui.selected;
-        lines.push(`- **UI needs**: ${selected}${ui.notes ? ` — ${ui.notes}` : ""}`);
-    }
-    const events = result.answers["events"];
-    if (events) {
-        const selected = Array.isArray(events.selected) ? events.selected[0] : events.selected;
-        lines.push(`- **Event hooks**: ${selected}${events.notes ? ` — ${events.notes}` : ""}`);
-    }
-    const persistence = result.answers["persistence"];
-    if (persistence) {
-        const selected = Array.isArray(persistence.selected) ? persistence.selected[0] : persistence.selected;
-        lines.push(`- **State persistence**: ${selected}${persistence.notes ? ` — ${persistence.notes}` : ""}`);
-    }
-    const complexity = result.answers["complexity"];
-    if (complexity) {
-        const selected = Array.isArray(complexity.selected) ? complexity.selected[0] : complexity.selected;
-        lines.push(`- **Complexity**: ${selected}${complexity.notes ? ` — ${complexity.notes}` : ""}`);
-    }
-    return lines.join("\n");
-}
-function sendPrompt(description, result, pi) {
-    const contextSection = `\n## Context gathered from user\n${formatAnswers(result)}\n`;
-    // Determine which doc sections to highlight based on answers
-    const uiAnswer = result.answers["ui"];
-    const uiSelected = uiAnswer
-        ? (Array.isArray(uiAnswer.selected) ? uiAnswer.selected[0] : uiAnswer.selected)
-        : "";
-    const eventsAnswer = result.answers["events"];
-    const eventsSelected = eventsAnswer
-        ? (Array.isArray(eventsAnswer.selected) ? eventsAnswer.selected[0] : eventsAnswer.selected)
-        : "";
-    const persistenceAnswer = result.answers["persistence"];
-    const persistenceSelected = persistenceAnswer
-        ? (Array.isArray(persistenceAnswer.selected) ? persistenceAnswer.selected[0] : persistenceAnswer.selected)
-        : "";
-    const docHints = [
-        "- `~/.gsd/agent/docs/extending-pi/01-what-are-extensions.md` — capabilities overview",
-        "- `~/.gsd/agent/docs/extending-pi/03-getting-started.md` — minimal extension, hot reload",
-        "- `~/.gsd/agent/docs/extending-pi/08-extensioncontext-what-you-can-access.md` — ExtensionContext API",
-        "- `~/.gsd/agent/docs/extending-pi/09-extensionapi-what-you-can-do.md` — ExtensionAPI: registration, messaging",
-        "- `~/.gsd/agent/docs/extending-pi/22-key-rules-gotchas.md` — must-read rules before shipping",
-    ];
-    if (uiSelected.includes("custom component")) {
-        docHints.push("- `~/.gsd/agent/docs/extending-pi/12-custom-ui-visual-components.md` — dialogs, widgets, overlays");
-        docHints.push("- `~/.gsd/agent/docs/pi-ui-tui/06-ctx-ui-custom-full-custom-components.md` — ctx.ui.custom() API");
-        docHints.push("- `~/.gsd/agent/docs/pi-ui-tui/07-built-in-components-the-building-blocks.md` — Text, Box, SelectList");
-        docHints.push("- `~/.gsd/agent/docs/pi-ui-tui/09-keyboard-input-how-to-handle-keys.md` — Key, matchesKey");
-        docHints.push("- `~/.gsd/agent/docs/pi-ui-tui/10-line-width-the-cardinal-rule.md` — truncation, width rules");
-        docHints.push("- `~/.gsd/agent/docs/pi-ui-tui/19-building-a-complete-component-step-by-step.md` — step-by-step guide");
-        docHints.push("- `~/.gsd/agent/docs/pi-ui-tui/21-common-mistakes-and-how-to-avoid-them.md` — pitfalls");
-    }
-    else if (uiSelected.includes("Dialogs")) {
-        docHints.push("- `~/.gsd/agent/docs/pi-ui-tui/04-built-in-dialog-methods.md` — select, confirm, input, editor");
-    }
-    else if (uiSelected.includes("Status")) {
-        docHints.push("- `~/.gsd/agent/docs/pi-ui-tui/05-persistent-ui-elements.md` — status, widgets, footer, header");
-    }
-    if (uiSelected.includes("tool") || result.answers["purpose"]) {
-        docHints.push("- `~/.gsd/agent/docs/extending-pi/14-custom-rendering-controlling-what-the-user-sees.md` — renderCall / renderResult");
-    }
-    if (eventsSelected && !eventsSelected.includes("standalone")) {
-        docHints.push("- `~/.gsd/agent/docs/extending-pi/07-events-the-nervous-system.md` — all events reference");
-    }
-    if (eventsSelected.includes("context / prompt")) {
-        docHints.push("- `~/.gsd/agent/docs/extending-pi/15-system-prompt-modification.md` — system prompt hooks");
-    }
-    if (persistenceSelected.includes("session")) {
-        docHints.push("- `~/.gsd/agent/docs/extending-pi/13-state-management-persistence.md` — pi.appendEntry, session state");
-    }
-    const prompt = `Create a new pi extension based on this description:
-
-"${description}"
-${contextSection}
-## Reference documentation
-
-Before writing any code, read the relevant docs below. They contain the exact APIs, rules, and patterns for building pi extensions — do not guess or rely on general TypeScript knowledge alone.
-
-${docHints.join("\n")}
-
-## Output
-
-Write the complete implementation as a single self-contained extension file:
-
-\`~/.gsd/agent/extensions/<kebab-case-name>.ts\`
-
-Then register it in the main extensions index:
-
-\`~/.gsd/agent/extensions/index.ts\` — import and call the new extension's default export alongside existing ones
-
-## Rules you must follow exactly
-
-- Extension entry point: \`export default function <camelCaseName>(pi: ExtensionAPI): void { ... }\`
-- Import type: \`import type { ExtensionAPI, ExtensionContext, ExtensionCommandContext } from "@gsd/pi-coding-agent";\`
-- \`pi\` is the registration surface — call \`pi.registerCommand\`, \`pi.registerTool\`, \`pi.on\`, \`pi.registerShortcut\` inside the default export
-- \`ctx\` (ExtensionCommandContext or ExtensionContext) is passed to handlers and event callbacks — never stored, never assumed available globally
-- To send a message to the agent: \`pi.sendUserMessage("...")\` or \`pi.sendMessage({ content, display }, { triggerTurn })\`
-- To show UI: \`ctx.ui.notify\`, \`ctx.ui.select\`, \`ctx.ui.input\`, \`ctx.ui.confirm\`, \`ctx.ui.custom\`
-- To run shell commands: \`await pi.exec("cmd", ["arg1"])\` — returns \`{ stdout, stderr, exitCode }\`
-- Events use \`pi.on("event_name", async (event, ctx) => { ... })\`
-- No direct file I/O without \`node:fs\` — import it explicitly if needed
-- Read the gotchas file before finalising: \`22-key-rules-gotchas.md\`
-
-After writing the files, run \`/reload\` to load the new extension.`;
-    pi.sendUserMessage(prompt);
-}