Integrate LLM API RAG fine-tuning prompt engineering พร้อมคุม token cost และกัน prompt injection

Blockchain/Web3 Developer — smart contract (Solidity/Rust), ERC-20/721/1155, DeFi protocol, security audit เบื้องต้น, deploy testnet/mainnet

เขียน SQL วิเคราะห์ insight สรุป finding และออกแบบ dashboard — แปลงตัวเลขเป็น action

ออกแบบ schema normalize indexing optimize query migration safety — Postgres MySQL MongoDB

พัฒนาเกมด้วย Unity C# หรือ Godot GDScript พร้อม physics save system UI pattern

สร้าง mobile app iOS/Android ด้วย React Native Flutter Swift Kotlin พร้อม deploy ขึ้น store

Review โค้ด debug refactor และอธิบายโค้ดแบบ senior developer รองรับทุกภาษา

สร้าง test case ครอบคลุม edge case พร้อม bug report template และ regression checklist

ตรวจช่องโหว่ OWASP Top 10 JWT SQL injection CSRF พร้อมเสนอวิธีแก้แบบ pentester

เขียน README API doc tutorial ADR ที่นักพัฒนาเข้าใจและใช้งานได้จริง

สร้าง concept design + mood board + Midjourney render prompt + material spec สำหรับงานสถาปัตยกรรม

ออกแบบภายใน — functional layout, furniture spec, lighting plan, material palette, mood board พร้อม render prompt

ผู้ช่วย internal audit — audit plan, sampling, test of controls, finding documentation, SOX/ISO 27001 checklist

ผู้จัดการร้านออนไลน์ — Shopee/Lazada/Shopify, conversion funnel, cart abandonment, fulfillment, return policy

ผู้ช่วยจัดซื้อมืออาชีพ — RFP/RFQ template, supplier scorecard, TCO, contract negotiation, e-procurement

AI Agile/Scrum Coach — sprint ceremony, velocity tracking, retrospective, team coaching รองรับทีม dev + non-dev

>

>

>

>

Generate 3D models, textures, images, rig characters, and animate them using the Meshy AI API. Handles API key detection, setup, and all generation workflows via direct HTTP calls. Use when the user asks to create 3D models, convert text/images to 3D, texture models, rig or animate characters, or interact with the Meshy API.

3D print models generated with Meshy AI. Handles slicer detection, white model printing, multi-color printing via API, and print-optimized download workflows. Use when the user mentions 3D printing, slicing, Bambu, OrcaSlicer, Prusa, Cura, Creality Print, Elegoo, Anycubic, multicolor, multi-color, 3mf, or wants to print a figurine, miniature, or physical model.

Render Compose @Preview functions to PNG outside Android Studio. Use this to verify UI changes, iterate on designs, and compare before/after states across Android (Jetpack Compose) and Compose Multiplatform Desktop projects.

Use when the user asks to commit, create a git commit, checkpoint changes, save work to git, or write a commit message.

Use when the goal is to reduce code size, remove unnecessary complexity, or simplify a module without removing user-facing behavior.

Use when generating code from a Pencil (.pen) design, syncing code to a .pen file, or aligning an implementation with .pen changes.

Use when creating, editing, or visually reviewing Pencil (.pen) screens, components, or flows.

Use when reviewing a pull request, merge request, or local diff for correctness, security, and code quality.

Use after making changes to run the project's quality gates and report pass or fail for each relevant check.

Use when you want a deeper multi-angle quality review before shipping, especially after substantial or risky changes.

Use when the user asks for a substantial refactor, structural cleanup, or deep codebase improvement beyond a small localized change.

Use when creating a new skill, rewriting an existing skill, or running evaluations to improve skill quality and trigger accuracy.

Use when the user needs a monthly work summary, time-tracking report, or commit-based activity recap.

Use when rewriting an existing skill so Opus 4.7 (or any frontier reasoning model) executes it with better judgment, less scaffolding, and tighter calibration. Do NOT trigger for new skills from scratch (use skill-creator) or for prompt artifacts (use prompt).

Use when reviewing a user-facing page or flow for usability issues, confusion, friction, or missed UX principles.

Use when removing a git worktree and cleaning up its associated local branch from the main repository.

forge-bugfix

> **在会话收尾时复盘，不要在中途插入。** 中途复盘的问题：工作未收敛导致判断被推翻、全量重新生成浪费 token、知识调研的 subagent 是最重的环节重跑成本最高。最佳节奏：一个会话的工作做完 → 复盘一次 → 结束。同会话迭代机制仅作为兜底方案保留。

forge-qa

forge-review

Generates conventional commit messages from staged diffs, enforcing semantic versioning conventions and team-defined scopes.

Discover and execute specialized skills from a curated marketplace catalog on GitHub, fetched on-demand and run ephemerally (no permanent installation). The catalog covers a broad and growing range — document generation (PDF/Word/Excel/PowerPoint), visual design, presentations, web app testing, animated GIFs, dev tooling (MCP scaffolding, frontend components), creative outputs (algorithmic art), technical documentation lookup (library/API/framework references), and more. Consult this skill whenever the user asks for a capability that could plausibly be covered by a community skill — the catalog is the source of truth, not this description.

Collective research memory — search what others have paid to learn, save your own findings, view contribution stats. Use when the user wants to check prior research before running a web search, save new findings, or view wellread stats — without the wellread MCP loaded.

Helps detection engineers identify coverage blind spots in system command execution detection rules, evaluate bypass risks, and reduce false negatives. Strictly limited to system command execution rules; not applicable to browser credential theft or abnormal process creation without command semantics.

'Frontend design blueprints with craftsmanship scoring, self-critique

cnb-cool-integration

'AI image generation and editing: e-commerce templates (hero/banner/detail/lifestyle),

ui-restructure

>-

>-