@hasna/skills 0.1.46 → 0.1.47

package/dist/index.js

@@ -170,7 +170,8 @@ var SKILL_ALIASES = {
   "pdf-reader": "read-pdf",
   "generate-image": "image",
   "image-generator": "image",
-  "create-blog-article": "blog-article"
+  "create-blog-article": "blog-article",
+  "skill-diff": "diff-viewer"
 };
 function normalizeSkillSlug(name) {
   return name.trim();
@@ -199,7 +200,7 @@ var PREMIUM_SKILLS = [
   { slug: "brand-kit", displayName: "Brand Kit", tier: "premium", costCents: 400, providers: ["hosted"], description: "Hosted brand kit with logo usage, palette, typography, brand voice, sample applications, Markdown guide, PDF guide, and SVG assets" },
   { slug: "generate-book-cover", displayName: "Book Cover", tier: "premium", costCents: 20, providers: ["gpt-image-2"], description: "Professional book cover design from title and genre" },
   { slug: "remove-background", displayName: "Remove Background", tier: "premium", costCents: 10, providers: ["gemini-3-pro"], description: "AI-powered background removal from images" },
-  { slug: "transcript", displayName: "Transcript", tier: "premium", costCents: 10, providers: ["whisper"], description: "Audio/video transcription with timestamps" },
+  { slug: "transcript", displayName: "Transcript", tier: "premium", costCents: 10, providers: ["openai", "elevenlabs", "deepgram", "hosted"], description: "Audio/video transcription with timestamps, diarization, and URL support" },
   { slug: "webcrawling", displayName: "Web Crawling", tier: "premium", costCents: 5, providers: ["firecrawl"], description: "Structured web page crawling and extraction" },
   { slug: "browse", displayName: "Browse", tier: "premium", costCents: 5, providers: ["browser"], description: "Web browsing and page interaction" },
   { slug: "read-pdf", displayName: "Read PDF", tier: "premium", costCents: 5, providers: ["cerebras"], description: "Hosted PDF extraction and structured content analysis" },
@@ -2806,9 +2807,9 @@ var MEDIA_PROCESSING_SKILLS = [
   {
     name: "transcript",
     displayName: "Transcript",
-    description: "Generate transcripts from audio and video files with timestamps",
+    description: "Transcribe audio, video, and media URLs with OpenAI GPT-4o, ElevenLabs Scribe v2, DeepGram, or hosted runtime",
     category: "Media Processing",
-    tags: ["transcript", "audio", "video", "speech-to-text"]
+    tags: ["transcript", "audio", "video", "speech-to-text", "diarization", "youtube"]
   },
   {
     name: "video-cut-suggester",
@@ -15482,7 +15483,7 @@ function finalize(ctx, schema) {
     result.$schema = "http://json-schema.org/draft-07/schema#";
   } else if (ctx.target === "draft-04") {
     result.$schema = "http://json-schema.org/draft-04/schema#";
-  } else if (ctx.target === "openapi-3.0") {}
+  } else if (ctx.target === "openapi-3.0") {} else {}
   if (ctx.external?.uri) {
     const id = ctx.external.registry.get(schema)?.id;
     if (!id)
@@ -15730,7 +15731,7 @@ var literalProcessor = (schema, ctx, json, _params) => {
     if (val === undefined) {
       if (ctx.unrepresentable === "throw") {
         throw new Error("Literal `undefined` cannot be represented in JSON Schema");
-      }
+      } else {}
     } else if (typeof val === "bigint") {
       if (ctx.unrepresentable === "throw") {
         throw new Error("BigInt literals cannot be represented in JSON Schema");
@@ -18811,7 +18812,7 @@ import { dirname as dirname4, relative as relative3 } from "path";
 // package.json
 var package_default = {
   name: "@hasna/skills",
-  version: "0.1.46",
+  version: "0.1.47",
   description: "Skills library for AI coding agents",
   type: "module",
   bin: {
@@ -18880,7 +18881,7 @@ var package_default = {
     typescript: "^5"
   },
   dependencies: {
-    "@hasna/events": "^0.1.3",
+    "@hasna/events": "^0.1.7",
     "@modelcontextprotocol/sdk": "^1.26.0",
     chalk: "^5.3.0",
     commander: "^12.1.0",
@@ -19064,18 +19065,29 @@ var runOutputSchema = objectSchema({
   exitCode: { type: "number", description: "Process exit code for local runs." },
   skill: stringSchema("Canonical skill slug."),
   remote: { type: "boolean", description: "Whether the skill was submitted to the hosted runtime." },
-  stdout: stringSchema("Captured stdout for local runs."),
-  stderr: stringSchema("Captured stderr for local runs."),
+  stdoutPreview: objectSchema({
+    text: stringSchema("Truncated stdout preview."),
+    length: { type: "number" },
+    truncated: { type: "boolean" }
+  }, [], "Default compact stdout preview."),
+  stderrPreview: objectSchema({
+    text: stringSchema("Truncated stderr preview."),
+    length: { type: "number" },
+    truncated: { type: "boolean" }
+  }, [], "Default compact stderr preview."),
+  stdout: stringSchema("Captured stdout for local runs when detail:true is requested."),
+  stderr: stringSchema("Captured stderr for local runs when detail:true is requested."),
   id: stringSchema("Remote run id when submitted remotely."),
   localRunId: stringSchema("Local run metadata id."),
   status: stringSchema("Run lifecycle status."),
   pricing: pricingSchema,
-  remoteRun: objectSchema({}, [], "Normalized hosted remote run contract.", true),
-  run: objectSchema({}, [], "Local run metadata.", true),
+  remoteRun: objectSchema({}, [], "Compact hosted remote run summary by default; full contract when detail:true is requested.", true),
+  run: objectSchema({}, [], "Compact local run metadata by default; full metadata when detail:true is requested.", true),
   nextActions: objectSchema({
     poll: stringSchema("Command to poll run status."),
     download: stringSchema("Command to download artifacts.")
-  })
+  }),
+  detailHint: stringSchema("How to request the complete payload.")
 }, [], "Skill run result.");
 var toolContracts = [
   {
@@ -19123,7 +19135,7 @@ var toolContracts = [
   {
     name: "list_skills",
     title: "List Skills",
-    description: "List skills from the basic or full registry profile.",
+    description: "List skills from the basic or full registry profile. Returns a compact paged envelope by default.",
     params: ["category?", "profile?", "detail?", "limit?", "offset?"],
     category: "discovery",
     sideEffects: "none",
@@ -19139,7 +19151,11 @@ var toolContracts = [
       skills: arraySchema(skillSummarySchema),
       total: { type: "number" },
       offset: { type: "number" },
-      limit: { type: "number" }
+      limit: { type: "number" },
+      nextOffset: { type: "number" },
+      hasMore: { type: "boolean" },
+      nextArguments: objectSchema({}, [], "Arguments for the next page.", true),
+      detailHint: stringSchema("How to request fuller skill objects.")
     })
   },
   {
@@ -19160,7 +19176,7 @@ var toolContracts = [
   {
     name: "search_skills",
     title: "Search Skills",
-    description: "Search skills by name, description, or tags.",
+    description: "Search skills by name, description, or tags. Returns a compact paged envelope by default.",
     params: ["query", "profile?", "detail?", "limit?", "offset?"],
     category: "discovery",
     sideEffects: "none",
@@ -19172,7 +19188,16 @@ var toolContracts = [
       limit: { type: "number", minimum: 0 },
       offset: { type: "number", minimum: 0 }
     }, ["query"]),
-    outputSchema: objectSchema({ skills: arraySchema(skillSummarySchema) })
+    outputSchema: objectSchema({
+      skills: arraySchema(skillSummarySchema),
+      total: { type: "number" },
+      offset: { type: "number" },
+      limit: { type: "number" },
+      nextOffset: { type: "number" },
+      hasMore: { type: "boolean" },
+      nextArguments: objectSchema({}, [], "Arguments for the next page.", true),
+      detailHint: stringSchema("How to request fuller skill objects.")
+    })
   },
   {
     name: "get_skill_info",
@@ -19289,8 +19314,8 @@ var toolContracts = [
   {
     name: "run_skill",
     title: "Run Skill",
-    description: "Run a skill locally or through a configured remote runner.",
-    params: ["name", "input?", "args?", "approved?"],
+    description: "Run a skill locally or through a configured remote runner. Returns compact stdout/stderr previews and run summaries by default; pass detail:true for full records.",
+    params: ["name", "input?", "args?", "approved?", "detail?"],
     category: "execution",
     sideEffects: "local-process-or-remote-run",
     stable: true,
@@ -19298,28 +19323,33 @@ var toolContracts = [
       name: skillNameInput,
       input: runInputSchema,
       args: runArgsSchema,
-      approved: paidRunApprovalSchema
+      approved: paidRunApprovalSchema,
+      detail: { type: "boolean", default: false, description: "Return full stdout/stderr, remote run, and local run metadata." }
     }, ["name"]),
     outputSchema: runOutputSchema
   },
   {
     name: "get_run_status",
     title: "Get Run Status",
-    description: "Fetch remote run status and next actions.",
-    params: ["run_id"],
+    description: "Fetch remote run status and next actions. Returns a compact status summary by default; pass detail:true for the complete remote run payload.",
+    params: ["run_id", "detail?"],
     category: "execution",
     sideEffects: "none",
     stable: true,
-    inputSchema: objectSchema({ run_id: stringSchema("Remote or local run id.") }, ["run_id"]),
+    inputSchema: objectSchema({
+      run_id: stringSchema("Remote or local run id."),
+      detail: { type: "boolean", default: false, description: "Return the complete remote run payload." }
+    }, ["run_id"]),
     outputSchema: objectSchema({
       contractVersion: { type: "number", description: "Remote run payload contract version." },
       runId: stringSchema("Remote run id."),
       localRunId: stringSchema("Local run id."),
-      run: objectSchema({}, [], "Normalized remote run status.", true),
+      run: objectSchema({}, [], "Compact remote run status by default; full status when detail:true is requested.", true),
       nextActions: objectSchema({
         poll: stringSchema("Command to poll run status."),
         download: stringSchema("Command to download artifacts.")
-      })
+      }),
+      detailHint: stringSchema("How to request the complete payload.")
     })
   },
   {
@@ -19414,13 +19444,25 @@ var toolContracts = [
   {
     name: "list_schedules",
     title: "List Schedules",
-    description: "List scheduled skill runs.",
-    params: [],
+    description: "List scheduled skill runs as a compact paged envelope.",
+    params: ["limit?", "offset?"],
     category: "scheduling",
     sideEffects: "none",
     stable: true,
-    inputSchema: objectSchema(),
-    outputSchema: arraySchema(objectSchema({}, [], "Schedule record.", true))
+    inputSchema: objectSchema({
+      limit: { type: "number", minimum: 0 },
+      offset: { type: "number", minimum: 0 }
+    }),
+    outputSchema: objectSchema({
+      schedules: arraySchema(objectSchema({}, [], "Compact schedule record.", true)),
+      total: { type: "number" },
+      offset: { type: "number" },
+      limit: { type: "number" },
+      nextOffset: { type: "number" },
+      hasMore: { type: "boolean" },
+      nextArguments: objectSchema({}, [], "Arguments for the next page.", true),
+      detailHint: stringSchema("How to request complete schedule details.")
+    })
   },
   {
     name: "remove_schedule",

package/dist/lib/compact-output.d.ts

@@ -0,0 +1,32 @@
+export declare const DEFAULT_LIST_LIMIT = 30;
+export declare const DEFAULT_SEARCH_LIMIT = 20;
+export declare const DEFAULT_TAG_LIMIT = 80;
+export declare const DEFAULT_MCP_LIMIT = 25;
+export declare const MAX_PAGE_LIMIT = 200;
+export declare const DEFAULT_PREVIEW_CHARS = 600;
+export interface Page<T> {
+    items: T[];
+    total: number;
+    offset: number;
+    limit: number;
+    hasMore: boolean;
+    nextOffset: number | null;
+}
+export declare function truncateText(value: unknown, maxChars?: number): string;
+export declare function previewText(value: unknown, maxChars?: number): {
+    text: string;
+    length: number;
+    truncated: boolean;
+};
+export declare function parsePageLimit(value: string | number | undefined, fallback: number, options?: {
+    max?: number;
+    allowAll?: boolean;
+}): number;
+export declare function parsePageOffset(value: string | number | undefined, fallback?: number): number;
+export declare function paginate<T>(items: T[], options: {
+    limit: number;
+    offset?: number;
+}): Page<T>;
+export declare function showingLabel(total: number, shown: number, offset: number): string;
+export declare function compactRunRecord(run: any): Record<string, unknown>;
+export declare function compactRemoteRun(run: any): Record<string, unknown>;

package/dist/lib/skill-aliases.d.ts

@@ -11,6 +11,7 @@ export declare const SKILL_ALIASES: {
     readonly "generate-image": "image";
     readonly "image-generator": "image";
     readonly "create-blog-article": "blog-article";
+    readonly "skill-diff": "diff-viewer";
 };
 export type SkillAlias = keyof typeof SKILL_ALIASES;
 export declare function normalizeSkillSlug(name: string): string;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@hasna/skills",
-  "version": "0.1.46",
+  "version": "0.1.47",
   "description": "Skills library for AI coding agents",
   "type": "module",
   "bin": {
@@ -69,7 +69,7 @@
     "typescript": "^5"
   },
   "dependencies": {
-    "@hasna/events": "^0.1.3",
+    "@hasna/events": "^0.1.7",
     "@modelcontextprotocol/sdk": "^1.26.0",
     "chalk": "^5.3.0",
     "commander": "^12.1.0",

package/skills/transcript/SKILL.md

@@ -1,100 +1,80 @@
 ---
 name: transcript
-description: Transcribe audio and video files using ElevenLabs Scribe, OpenAI Whisper, or Google Gemini. Supports automatic chunking for large files, speaker diarization, timestamps, and multiple output formats (text, SRT, VTT, JSON).
+description: Transcribe audio, video, YouTube, Vimeo, and generic media URLs with iapp-transcriber or the hosted Skills runtime. Supports OpenAI GPT-4o transcription, OpenAI diarization, ElevenLabs Scribe v2, DeepGram, chunking, source metadata, subtitles, and JSON outputs.
 ---
 
-# Audio Transcription Skill
+# Transcript
 
-This skill provides high-quality speech-to-text transcription using multiple AI providers. It automatically handles large files through compression and chunking.
+Create transcripts from local audio/video files or media URLs. Use this skill when the user asks to transcribe, caption, diarize, summarize, or package spoken audio/video content.
 
-This CLI is API-backed. Set `SKILLS_API_KEY` when routing through the hosted skills/connectors runtime; provider-specific keys are managed by that runtime.
+## Choose The Runtime
 
-## Supported Providers
+- Use the hosted Skills runtime when the user explicitly runs `skills run transcript`, needs remote execution, or has only `SKILLS_API_KEY` configured.
+- Use local `iapp-transcriber` when you are on this machine and need direct access to local files, YouTube/Vimeo/generic `yt-dlp` sources, transcript DB records, MCP tools, comments, exports, or OpenLoops follow-up workflows.
+- The local command is `transcriber` when installed, or `bun run src/cli/index.ts` from `/home/hasna/Workspace/hasnaxyz/internalapp/iapp-transcriber`.
 
-### ElevenLabs Scribe
-- **Accuracy**: 96.7% for English (industry-leading)
-- **Max file size**: 3GB / 10 hours
-- **Features**: Speaker diarization (up to 32 speakers), word-level timestamps
-- **Cost**: $0.40/hour
-- **Best for**: Multi-speaker recordings, highest accuracy needs
+## Hosted Usage
 
-### OpenAI Whisper
-- **Accuracy**: Excellent
-- **Max file size**: 25MB (automatic chunking for larger files)
-- **Features**: Segment timestamps, language detection
-- **Cost**: $0.006/min ($0.003/min with GPT-4o Mini)
-- **Best for**: Standard transcription, good balance of cost and quality
+```bash
+skills run transcript --source ./meeting.mp3 --title "Design review" --provider openai
+skills run transcribe --source https://www.youtube.com/watch?v=... --provider openai --diarize
+```
 
-### Google Gemini
-- **Accuracy**: Very good
-- **Max file size**: 2GB
-- **Features**: Multimodal analysis, summarization capabilities
-- **Cost**: ~$0.09-0.23/hour (generous free tier available)
-- **Best for**: Cost-sensitive projects, multimodal needs
+Poll hosted runs with `skills runs status <run-id>` and download outputs with `skills exports download <run-id>`.
 
-## Usage
+## Local Usage
 
-### Basic Transcription
 ```bash
-bun run src/index.ts transcribe \
-  --provider openai \
-  --input ./recording.mp3
+transcriber transcribe ./meeting.mp3 --provider openai --json
+transcriber transcribe https://www.youtube.com/watch?v=... --provider openai --model gpt-4o-transcribe --json
+transcriber transcribe ./meeting.mp3 --provider openai --diarize --json
+transcriber export <transcript-id> --format srt --output captions.srt
 ```
 
-### With Speaker Diarization
-```bash
-bun run src/index.ts transcribe \
-  --provider elevenlabs \
-  --input ./meeting.mp3 \
-  --diarize \
-  --timestamps \
-  --format srt
-```
+Local provider defaults:
 
-### Export to Subtitles
-```bash
-bun run src/index.ts transcribe \
-  --provider gemini \
-  --input ./video.mp4 \
-  --format vtt \
-  --output ./captions.vtt
-```
+- `openai`: default, uses `gpt-4o-transcribe`; `--diarize` uses `gpt-4o-transcribe-diarize`.
+- `elevenlabs`: uses `scribe_v2`, supports diarization and keyterms.
+- `deepgram`: uses Nova-3, supports diarization.
 
-### View Provider Info
-```bash
-bun run src/index.ts providers
-```
+Local requirements:
 
-## Output Formats
+- A configured provider credential for the selected local provider.
+- `yt-dlp` for remote media URLs. Set `YTDLP_PATH` if needed.
+- `ffmpeg`/`ffprobe`; the local app bundles npm ffmpeg/ffprobe and also respects `FFMPEG_PATH` and `FFPROBE_PATH`.
 
-| Format | Extension | Description |
-|--------|-----------|-------------|
-| text | .txt | Plain text transcript |
-| srt | .srt | SubRip subtitle format |
-| vtt | .vtt | WebVTT subtitle format |
-| json | .json | Full structured data with metadata |
+## Workflow
 
-## Large File Handling
+1. Inspect source metadata first for URLs:
 
-The skill automatically handles files larger than provider limits:
+   ```bash
+   transcriber info <url> --json
+   ```
 
-- **Compression**: For OpenAI, files are first compressed using Opus codec
-- **Chunking**: Files are split into 10-minute segments with overlap
-- **Merging**: Results are intelligently merged to avoid duplicates
+2. Download audio when the user asks to keep media:
 
-## Configuration
+   ```bash
+   transcriber download <url> --format mp3 --json
+   ```
 
-```bash
-export SKILLS_API_KEY=your_skill_api_key
-```
+3. Transcribe with JSON for automation:
 
-## Dependencies
+   ```bash
+   transcriber transcribe <path-or-url> --provider openai --json
+   ```
 
-For chunking support (OpenAI with large files):
-- `ffmpeg` - Audio processing
-- `ffprobe` - Duration detection
+4. Export or post-process:
 
-Install on macOS:
-```bash
-brew install ffmpeg
-```
+   ```bash
+   transcriber get <id> --json
+   transcriber export <id> --format txt --output transcript.txt
+   transcriber summarize <id>
+   ```
+
+5. For repeat work, create OpenLoops command loops around JSON-producing commands, for example `transcriber feed check --json --dry-run`.
+
+## Safety
+
+- Only fetch URLs the user is authorized to process.
+- The local app rejects private/local URL hosts by default; set `TRANSCRIBER_ALLOW_PRIVATE_URLS=1` only for trusted internal sources.
+- Prefer `--json` for scripts and OpenLoops so failures include a structured transcript record and nonzero exit code.

package/skills/transcript/package.json

@@ -2,7 +2,7 @@
   "name": "transcript",
   "version": "1.0.0",
   "private": true,
-  "description": "Audio/video transcription skill supporting ElevenLabs, OpenAI Whisper, and Google Gemini with automatic chunking for large files",
+  "description": "Audio/video transcription skill backed by iapp-transcriber, OpenAI GPT-4o transcription, ElevenLabs Scribe v2, and hosted Skills runtime",
   "type": "module",
   "skills": {
     "runtime": "hosted",

package/skills/apidocs/.claude/settings.json

@@ -1,5 +0,0 @@
-{
-  "env": {
-    "CLAUDE_CODE_TASK_LIST_ID": "service-apidocs-dev"
-  }
-}

package/skills/hook/bunfig.toml

@@ -1,5 +0,0 @@
-[install]
-peer = false
-
-[run]
-bun = true

package/skills/implementation/bunfig.toml

@@ -1,5 +0,0 @@
-[install]
-peer = false
-
-[run]
-bun = true