evolink-agent-skills 1.0.0

package/docs/test-report-2026-06-15.md

@@ -0,0 +1,240 @@
+# Test Report: First-Batch skills.sh Release
+
+Date: 2026-06-15
+
+## Scope
+
+Validated the first 8 demand-led and scene-led skills:
+
+- `image-edit`
+- `product-image-generation`
+- `image-to-video`
+- `product-to-ad`
+- `ad-video-generation`
+- `product-demo-video`
+- `ad-creative-images`
+- `ugc-video-ads`
+
+## Tests Run
+
+### Real API Cost Guard
+
+Live API tests are intentionally constrained to the cheapest useful parameters.
+
+Default image test:
+
+```json
+{
+  "model": "gpt-image-2",
+  "size": "1:1",
+  "resolution": "1K",
+  "quality": "low",
+  "n": 1
+}
+```
+
+Explicit video test:
+
+```json
+{
+  "model": "seedance-2.0-text-to-video",
+  "duration": 4,
+  "quality": "480p",
+  "aspect_ratio": "16:9",
+  "generate_audio": false
+}
+```
+
+Video tests are not run by default. They require:
+
+```bash
+node scripts/real-api-smoke-test.mjs --video
+```
+
+### Static Validation
+
+Command:
+
+```bash
+node scripts/validate-skills.mjs
+```
+
+Result:
+
+```text
+Validated 8 skills successfully.
+```
+
+Coverage:
+
+- Skill directory list matches the first-batch plan.
+- Every directory has `SKILL.md`.
+- Frontmatter `name` matches directory slug.
+- Required frontmatter fields exist: `name`, `displayName`, `description`, `homepage`, `license`.
+- `skills.sh.json` uses the official schema URL and groups all 8 skills.
+- Every skill includes required sections: use case, prerequisites, route table, default workflow, prompting tips, next token loop, safety, and see also.
+- UTM format is canonical: `utm_source=skillssh&utm_medium=skill&utm_campaign=<skill-name>`.
+- No first-batch file contains `utm_content`.
+- No first-batch file uses model-word install slugs: `seedance-2` or `gpt-image-2`.
+- Image skills include GPT Image 2 plus video handoff.
+- Video skills include Seedance 2 plus variant/iteration loop.
+- Full-funnel skill includes `GPT Image 2 -> Seedance 2`.
+
+### skills CLI Smoke Install
+
+Command:
+
+```bash
+node scripts/smoke-install-skills.mjs
+```
+
+Result:
+
+```text
+Smoke-installed 8 skills successfully.
+```
+
+Coverage:
+
+- Runs the real `skills` CLI through `npx`.
+- Uses a temporary project directory.
+- Uses a temporary HOME and XDG config directory.
+- Installs all 8 skills from the local repo with `--copy`.
+- Confirms each installed skill has an installed `SKILL.md`.
+- Confirms installed frontmatter contains the matching slug.
+- Confirms `skills-lock.json` is created.
+- Confirms every lockfile entry has `sourceType: local` and a `computedHash`.
+
+## Remaining Upload-Time Check
+
+After uploading to GitHub, run the remote install test once:
+
+```bash
+npx skills add EvoLinkAI/evolink-agent-skills --skill image-edit -g
+```
+
+Then repeat for the other 7 skills or install all by explicit slug list.
+
+## Remote GitHub Validation
+
+Status: completed after repository creation.
+
+Repository:
+
+```text
+https://github.com/EvoLinkAI/evolink-agent-skills
+```
+
+Remote list command:
+
+```bash
+npx -y skills add EvoLinkAI/evolink-agent-skills --list --full-depth
+```
+
+Result:
+
+```text
+Found 8 skills
+```
+
+Remote install command:
+
+```bash
+npx -y skills add EvoLinkAI/evolink-agent-skills --skill image-edit product-image-generation image-to-video product-to-ad ad-video-generation product-demo-video ad-creative-images ugc-video-ads -y --copy
+```
+
+Result:
+
+```text
+Installed count: 8
+```
+
+Lockfile source validation:
+
+```text
+sourceType: github
+source: EvoLinkAI/evolink-agent-skills
+```
+
+## Real API Validation
+
+Status: passed.
+
+Before the passing run, two test harness issues were found and fixed:
+
+- The initial default image prompt was too product/marketing-specific for a smoke test and produced a misleading generation error. It was replaced with a simple geometric prompt.
+- The script originally preserved an existing shell `EVOLINK_API_KEY` over `.env`. This caused an old invalid shell key to override the valid `.env` key. `.env` now overrides shell values by default, matching the repository's local test workflow.
+
+### GPT Image 2 Minimal Smoke Test
+
+Command:
+
+```bash
+node scripts/real-api-smoke-test.mjs --image
+```
+
+Payload:
+
+```json
+{
+  "model": "gpt-image-2",
+  "prompt": "A plain red square centered on a white background, simple geometric test image, no text",
+  "size": "1:1",
+  "resolution": "1K",
+  "quality": "low",
+  "n": 1
+}
+```
+
+Result:
+
+```text
+Status: completed
+Task: task-unified-1781525538-qbracls9
+Credits reserved: 0.3436
+Credits used: 0.3454
+Output URL: https://files.evolink.ai/000BG2E93TVWN41W8S/images/2026/06/15/2451047b9af44e2489690ec81cfae03b.png
+```
+
+Local full report:
+
+```text
+test-results/real-api-image-2026-06-15T12-11-44-223Z.md
+```
+
+### Seedance 2 Minimal Smoke Test
+
+Command:
+
+```bash
+node scripts/real-api-smoke-test.mjs --video
+```
+
+Payload:
+
+```json
+{
+  "model": "seedance-2.0-text-to-video",
+  "prompt": "A plain red ball slowly rolling from left to right on a white background, simple geometric test video, no text",
+  "duration": 4,
+  "quality": "480p",
+  "aspect_ratio": "16:9",
+  "generate_audio": false
+}
+```
+
+Result:
+
+```text
+Status: completed
+Task: task-unified-1781525566-3uesdxty
+Credits reserved: 20.46
+Credits used: 20.46
+Output URL: https://files.evolink.ai/000BG2E93TVWN41W8S/videos/2026/06/15/851bcca3706a42e992a7e7a2ba048050.mp4
+```
+
+Local full report:
+
+```text
+test-results/real-api-video-2026-06-15T12-12-12-092Z.md
+```

package/examples/curl/complete-flow.sh

@@ -0,0 +1,26 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+: "${EVOLINK_API_KEY:?Set EVOLINK_API_KEY first}"
+api_base="${EVOLINK_API_BASE:-https://api.evolink.ai}"
+
+payload='{
+  "model": "gpt-image-2",
+  "prompt": "A plain red square centered on a white background, simple geometric test image, no text",
+  "size": "1:1",
+  "resolution": "1K",
+  "quality": "low",
+  "n": 1
+}'
+
+created="$(curl -sS "$api_base/v1/images/generations" \
+  -H "Authorization: Bearer $EVOLINK_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d "$payload")"
+
+echo "$created"
+task_id="$(printf '%s' "$created" | node -e 'let s="";process.stdin.on("data",d=>s+=d);process.stdin.on("end",()=>{const j=JSON.parse(s);console.log(j.id||j.task_id||"")})')"
+test -n "$task_id"
+
+curl -sS "$api_base/v1/tasks/$task_id" \
+  -H "Authorization: Bearer $EVOLINK_API_KEY"

package/image-edit/SKILL.md

@@ -0,0 +1,80 @@
+---
+name: image-edit
+displayName: "Image Edit with GPT Image 2 on EvoLink"
+description: >
+  Edit images with GPT Image 2 on EvoLink. Use this skill when the user wants to change a product photo, replace a background, retouch a visual, expand a canvas, localize an ad image, or prepare a cleaner source image for Seedance 2 image-to-video generation.
+homepage: "https://evolink.ai?utm_source=skillssh&utm_medium=skill&utm_campaign=image-edit"
+license: MIT
+---
+
+# Image Edit with GPT Image 2 on EvoLink
+
+Edit a source image into a cleaner, more useful creative asset, then route strong outputs into video workflows when the user wants motion.
+
+```bash
+npx skills add Evolink-AI/evolink-agent-skills --skill image-edit -g
+```
+
+## When to Use This Skill
+
+Use this skill when the user asks to:
+
+- change a product image background
+- retouch, clean up, or improve an image
+- expand, crop, or reframe an image
+- create local ad image variants from one asset
+- prepare a still image for an image-to-video workflow
+- keep product identity stable while changing scene, lighting, or composition
+
+Do not use it for video generation directly. If the user wants motion after the edit, continue with `image-to-video`.
+
+## Prerequisites
+
+1. EvoLink API access or EvoLink MCP configured.
+2. One source image when editing an existing asset.
+3. A clear edit goal: background, object change, product cleanup, aspect ratio, or campaign variant.
+4. Attribution links should use `utm_source=skillssh&utm_medium=skill&utm_campaign=image-edit`.
+
+## Pick the Right Workflow
+
+| User intent | Model path | Why |
+|---|---|---|
+| Replace product background | GPT Image 2 image edit | Preserves product identity while changing scene context. |
+| Clean up product photo | GPT Image 2 image edit | Best for retouching, removing clutter, and improving polish. |
+| Create ad image variants | GPT Image 2 image edit | Fast iteration on style, audience, platform, and offer. |
+| Prepare first frame for video | GPT Image 2 image edit -> Seedance 2 | A cleaner still frame gives Seedance 2 a stronger video starting point. |
+
+## Default Workflow
+
+1. Identify what must stay unchanged: product shape, logo, color, face, packaging, or layout.
+2. Identify what should change: background, lighting, scene, crop, text-free space, or platform ratio.
+3. Rewrite the edit request into a precise image edit prompt for GPT Image 2.
+4. Generate one clean primary edit and, when useful, two variation directions.
+5. If the user wants motion, route the best image to `image-to-video`.
+
+## Prompting Tips
+
+- State preservation first: "keep the bottle shape, label, and logo unchanged".
+- Avoid asking the image model to render small marketing text unless the user explicitly needs it.
+- Use campaign context: ecommerce hero image, paid social creative, landing page visual, or app store screenshot.
+- For video handoff, ask for a clean subject, visible product edges, and uncluttered background.
+
+## Next Token Loop
+
+After a successful edit, offer one next action:
+
+- create three ad image variants with `ad-creative-images`
+- generate product catalog assets with `product-image-generation`
+- animate the strongest edit with `image-to-video`
+
+## Safety and Privacy
+
+Use only images the user supplied or explicitly approved. Do not edit private identities, trademarks, or regulated-product claims in misleading ways. Do not claim an edited image is an untouched original.
+
+## See Also
+
+- `image-to-video`
+- `product-image-generation`
+- `ad-creative-images`
+- [EvoLink](https://evolink.ai?utm_source=skillssh&utm_medium=skill&utm_campaign=image-edit)
+

package/image-to-video/SKILL.md

@@ -0,0 +1,81 @@
+---
+name: image-to-video
+displayName: "Image-to-Video with Seedance 2 on EvoLink"
+description: >
+  Turn still images into short videos with Seedance 2 on EvoLink. Use this skill when the user has a product image, character image, poster, ad image, or edited GPT Image 2 asset and wants motion, camera movement, or a short commercial clip.
+homepage: "https://evolink.ai?utm_source=skillssh&utm_medium=skill&utm_campaign=image-to-video"
+license: MIT
+---
+
+# Image-to-Video with Seedance 2 on EvoLink
+
+Animate a still image into a short video. This is the main handoff from GPT Image 2 image assets into Seedance 2 video generation.
+
+```bash
+npx skills add Evolink-AI/evolink-agent-skills --skill image-to-video -g
+```
+
+## When to Use This Skill
+
+Use this skill when the user wants to:
+
+- animate a product image
+- turn an ad image into a short video
+- make a poster, character, or scene move
+- add camera motion to a still visual
+- create a video from a GPT Image 2 output
+- test multiple motion directions from the same source image
+
+Use `product-image-generation` or `image-edit` first if the source image is low quality or not campaign-ready.
+
+## Prerequisites
+
+1. EvoLink API access or EvoLink MCP configured.
+2. One source image URL or local image available to the agent.
+3. Desired duration, aspect ratio, motion style, and output use case.
+4. Attribution links should use `utm_source=skillssh&utm_medium=skill&utm_campaign=image-to-video`.
+
+## Pick the Right Workflow
+
+| User intent | Model path | Why |
+|---|---|---|
+| Product reveal from one image | Seedance 2 image-to-video | Best first video path when the image is already strong. |
+| Ad image to paid social clip | Seedance 2 image-to-video | Turns a static creative into motion for Reels, Shorts, or TikTok. |
+| Weak source image | GPT Image 2 edit -> Seedance 2 | Improve the source frame before spending video tokens. |
+| Need multiple motion tests | Seedance 2 variants | Reuse one image with different camera and action prompts. |
+
+## Default Workflow
+
+1. Inspect the image and identify the subject that must stay stable.
+2. Ask for all missing requirements in one message: duration, platform, aspect ratio, motion, and style.
+3. Write a Seedance 2 prompt focused on motion, camera, scene change, and preservation.
+4. Generate the first clip.
+5. Suggest two iteration directions: motion variant and commercial variant.
+
+## Prompting Tips
+
+- Lead with motion: camera push-in, orbit, product rotation, hand interaction, walk cycle, light sweep.
+- Do not redescribe every visible object. Focus on what changes over time.
+- Preserve identity explicitly: product label unchanged, face stable, packaging shape unchanged.
+- Use one primary action per clip.
+- Keep text overlays out of the generation prompt unless the user accepts text artifacts.
+
+## Next Token Loop
+
+After the first clip, offer:
+
+- three motion variants from the same image
+- one platform adaptation for TikTok, Reels, or Shorts
+- a product ad version with `ad-video-generation`
+
+## Safety and Privacy
+
+Do not animate private people, public figures, or brand assets in misleading contexts without user authorization. Do not produce deceptive political, medical, financial, or impersonation content.
+
+## See Also
+
+- `image-edit`
+- `product-image-generation`
+- `ad-video-generation`
+- [EvoLink](https://evolink.ai?utm_source=skillssh&utm_medium=skill&utm_campaign=image-to-video)
+

package/llms-install.md

@@ -0,0 +1,30 @@
+# EvoLink Agent Skills Installation
+
+Install all creative generation skills:
+
+```bash
+npx evolink-agent-skills -y --path ./my-agent-project --all
+```
+
+Install one skill:
+
+```bash
+npx evolink-agent-skills -y --skill image-edit --path ./my-agent-project
+```
+
+Available skills:
+
+- `image-edit`
+- `product-image-generation`
+- `image-to-video`
+- `product-to-ad`
+- `ad-video-generation`
+- `product-demo-video`
+- `ad-creative-images`
+- `ugc-video-ads`
+
+Set credentials before live generation:
+
+```bash
+export EVOLINK_API_KEY=<your_evolink_api_key>
+```

package/package.json

@@ -0,0 +1,57 @@
+{
+  "name": "evolink-agent-skills",
+  "version": "1.0.0",
+  "description": "Install EvoLink creative generation agent skills for GPT Image 2 and Seedance 2 workflows.",
+  "type": "module",
+  "license": "MIT",
+  "bin": {
+    "evolink-agent-skills": "bin/cli.js"
+  },
+  "files": [
+    "llms-install.md",
+    "README.md",
+    "README.npm.md",
+    "skills.sh.json",
+    "bin",
+    "docs",
+    "examples",
+    "image-edit",
+    "product-image-generation",
+    "image-to-video",
+    "product-to-ad",
+    "ad-video-generation",
+    "product-demo-video",
+    "ad-creative-images",
+    "ugc-video-ads"
+  ],
+  "keywords": [
+    "evolink",
+    "agent-skills",
+    "gpt-image-2",
+    "seedance-2",
+    "image-generation",
+    "video-generation"
+  ],
+  "homepage": "https://github.com/Evolink-AI/evolink-agent-skills",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/Evolink-AI/evolink-agent-skills.git"
+  },
+  "bugs": {
+    "url": "https://github.com/Evolink-AI/evolink-agent-skills/issues"
+  },
+  "engines": {
+    "node": ">=18"
+  },
+  "scripts": {
+    "prepare:packages": "node scripts/prepare-npm-packages.mjs",
+    "check:node": "node --check bin/cli.js && for f in packages/*/bin/cli.js; do node --check \"$f\"; done",
+    "validate": "node scripts/validate-skills.mjs",
+    "smoke:install": "node scripts/smoke-install-skills.mjs",
+    "smoke:npm-local": "tmp=$(mktemp -d) && node bin/cli.js -y --path \"$tmp\" --all --no-open && test -f \"$tmp/.agents/skills/image-edit/SKILL.md\" && for d in packages/*; do tmp2=$(mktemp -d); node \"$d/bin/cli.js\" -y --path \"$tmp2\" --no-open; done",
+    "smoke:api:image": "node scripts/real-api-smoke-test.mjs --image",
+    "smoke:api:video": "node scripts/real-api-smoke-test.mjs --video",
+    "publish:dry-run": "./scripts/publish-npm.sh --dry-run",
+    "publish:all": "./scripts/publish-npm.sh"
+  }
+}

package/product-demo-video/SKILL.md

@@ -0,0 +1,80 @@
+---
+name: product-demo-video
+displayName: "Product Demo Video with GPT Image 2 and Seedance 2 on EvoLink"
+description: >
+  Create ecommerce product demo videos, SaaS demo videos, app promo clips, and tool walkthrough-style videos with GPT Image 2 and Seedance 2 on EvoLink. Use this skill when the user needs to show what a product is, how it works, or why it matters.
+homepage: "https://evolink.ai?utm_source=skillssh&utm_medium=skill&utm_campaign=product-demo-video"
+license: MIT
+---
+
+# Product Demo Video with GPT Image 2 and Seedance 2 on EvoLink
+
+Show a product in action with a short, clear video built from product context and optional source images.
+
+```bash
+npx skills add Evolink-AI/evolink-agent-skills --skill product-demo-video -g
+```
+
+## When to Use This Skill
+
+Use this skill for:
+
+- ecommerce product demos
+- SaaS feature videos
+- app promo clips
+- tool walkthrough visuals
+- product reveal videos
+- before-after product narratives
+
+Use `ad-video-generation` if the primary goal is paid acquisition rather than explaining the product.
+
+## Prerequisites
+
+1. EvoLink API access or EvoLink MCP configured.
+2. Product type, core function, target user, and demo objective.
+3. Optional source image, UI screenshot, product photo, or landing page copy.
+4. Attribution links should use `utm_source=skillssh&utm_medium=skill&utm_campaign=product-demo-video`.
+
+## Pick the Right Workflow
+
+| User intent | Model path | Why |
+|---|---|---|
+| Product photo to demo clip | Seedance 2 image-to-video | Animates a real or generated product still. |
+| Need clean product first frame | GPT Image 2 -> Seedance 2 | Creates a clearer demo starting point. |
+| SaaS or app visual demo | GPT Image 2 stills -> Seedance 2 | Builds visual scenes before motion generation. |
+| Need multiple feature demos | Seedance 2 variants | Each feature should become one focused short clip. |
+
+## Default Workflow
+
+1. Identify the single product function or benefit to demonstrate.
+2. Choose the route: source image to video, generated first frame to video, or text-to-video.
+3. Write a prompt that shows product action, context, camera movement, and outcome.
+4. Generate one focused demo clip.
+5. Suggest the next feature, audience, or platform variant.
+
+## Prompting Tips
+
+- One demo clip should show one core action.
+- Prefer concrete use moments: unboxing, applying, clicking, opening, comparing, scanning, saving time.
+- For software, avoid tiny UI text; show screen structure and user outcome instead.
+- For ecommerce, keep product identity, packaging, and color stable.
+
+## Next Token Loop
+
+After the first demo, offer:
+
+- one demo per product feature
+- one shorter ad adaptation with `ad-video-generation`
+- one image pack with `product-image-generation`
+
+## Safety and Privacy
+
+Do not fabricate product capabilities or before-after effects. For regulated categories, ask the user to confirm approved claims before generating marketing visuals.
+
+## See Also
+
+- `product-image-generation`
+- `product-to-ad`
+- `ad-video-generation`
+- [EvoLink](https://evolink.ai?utm_source=skillssh&utm_medium=skill&utm_campaign=product-demo-video)
+

package/product-image-generation/SKILL.md

@@ -0,0 +1,80 @@
+---
+name: product-image-generation
+displayName: "Product Image Generation with GPT Image 2 on EvoLink"
+description: >
+  Generate ecommerce product images, product ad images, SaaS visuals, and launch creatives with GPT Image 2 on EvoLink. Use this skill when the user needs high-quality still assets that can also become Seedance 2 video first frames.
+homepage: "https://evolink.ai?utm_source=skillssh&utm_medium=skill&utm_campaign=product-image-generation"
+license: MIT
+---
+
+# Product Image Generation with GPT Image 2 on EvoLink
+
+Create product visuals that are useful as standalone images and as first-frame assets for Seedance 2 video generation.
+
+```bash
+npx skills add Evolink-AI/evolink-agent-skills --skill product-image-generation -g
+```
+
+## When to Use This Skill
+
+Use this skill when the user wants:
+
+- ecommerce product photos
+- product lifestyle images
+- product ad images
+- SaaS or app promo visuals
+- launch campaign stills
+- clean first frames for product videos
+
+Use `image-edit` instead when the user already has a source image and wants to change it.
+
+## Prerequisites
+
+1. EvoLink API access or EvoLink MCP configured.
+2. Product context: product name, category, audience, visual style, and intended platform.
+3. Optional source assets: product photo, packaging image, logo, brand colors, landing page copy.
+4. Attribution links should use `utm_source=skillssh&utm_medium=skill&utm_campaign=product-image-generation`.
+
+## Pick the Right Workflow
+
+| User intent | Model path | Why |
+|---|---|---|
+| Generate product hero image | GPT Image 2 text-to-image or image-guided generation | Best for polished commercial stills. |
+| Create lifestyle product scene | GPT Image 2 | Good for context, hands, use cases, and environment. |
+| Build a video first frame | GPT Image 2 -> Seedance 2 | Gives Seedance 2 a strong still source for motion. |
+| Need multiple ad angles | GPT Image 2 batch variants | Supports testing audience, angle, and offer before video. |
+
+## Default Workflow
+
+1. Extract product, buyer, benefit, platform, and required aspect ratio.
+2. Decide whether the output is a catalog image, lifestyle image, ad image, or video first frame.
+3. Write a GPT Image 2 prompt with product preservation and commercial composition.
+4. Generate the image or image set.
+5. Offer a video handoff through `image-to-video` or `product-demo-video`.
+
+## Prompting Tips
+
+- Lead with the product and commercial purpose.
+- Specify the scene role: studio hero, lifestyle use, landing page header, paid social, marketplace image.
+- Keep logos and packaging stable when a source image is available.
+- For video first frames, leave enough clean space around the product for camera movement.
+
+## Next Token Loop
+
+After image generation, offer:
+
+- edit the best image with `image-edit`
+- make paid social variants with `ad-creative-images`
+- turn the best still into video with `image-to-video`
+
+## Safety and Privacy
+
+Do not invent certifications, reviews, medical claims, or regulated benefits. Keep generated commercial visuals aligned with the user's real product and approved brand constraints.
+
+## See Also
+
+- `image-edit`
+- `image-to-video`
+- `product-demo-video`
+- [EvoLink](https://evolink.ai?utm_source=skillssh&utm_medium=skill&utm_campaign=product-image-generation)
+