discussion-limitations-review
Check interpretation quality, limitations, and overclaiming.
Check interpretation quality, limitations, and overclaiming.
Draft and fill SaaS agreement templates — cloud contract, MSA, order form, software license, pilot agreement, design partner agreement. Includes variants with SLAs and AI terms. Produces signable DOCX from Common Paper standard forms. Use when user says "SaaS agreement," "cloud contract," "MSA," "order form," "software license," "pilot agreement," or "design partner agreement."
Draft client-facing emails for legal services — cover notes for contract deliverables, redline summaries, deal status updates, and follow-ups. Use when composing or revising outbound emails to clients about legal work product. Triggers on "draft reply," "email to client," "cover note," "write back to," or any outbound email accompanying a legal deliverable.
Mailchimp's UI design system. Use when building interfaces inspired by Mailchimp's aesthetic - light mode, Inter font, 4px grid.
Generate or edit images using AI models (FLUX, Gemini). Use for general-purpose image generation including photos, illustrations, artwork, visual assets, concept art, and any image that isn't a technical diagram or schematic. For flowcharts, circuits, pathways, and technical diagrams, use the scientific-schematics skill instead.
Generate 5-view multi-angle images (front/side/back/top/45-degree) from one or many input photos using the Qwen Image multiple-angles 3D camera Gradio app (Hugging Face or ModelScope mirrors). Use when user asks to create multi-angle turnarounds, camera angle variants, or batch process 10+ images into consistent view sets and return a zip. Prefer API-based Gradio calls (no browser automation) with retries, stable filenames, and quota/queue handling.
Smart video editor for talking-head / oral-broadcast (口播) videos. Use when user sends a video file and asks to cut, trim, edit, or clean up a talking-head video. Handles: silence removal, subtitle burn-in, loudness normalization, jump-cut assembly. Default mode: silence-only cutting (simple and reliable). Filler-word cutting available via --filler-words flag but disabled by default. Supports Chinese and English speech. NOT for VFX, color grading, or non-speech video editing.
Generate high-quality Telegram voice notes (bubbles) using Microsoft Edge TTS and FFmpeg. Use when asked to send audio as a voice message, or when the user requires voice-bar replies instead of audio files. Handles edge-tts generation and OGG/Opus conversion for native Telegram voice compatibility.
Post-processing visual effects including chromatic aberration, vignette, depth of field, film grain, color grading, and LUT support. Use when adding cinematic polish, retro aesthetics, camera simulation, or atmospheric effects to 3D scenes. Essential for mood, style, and visual storytelling.
Complete FFmpeg toolkit for video/audio manipulation, analysis, and production. Covers probing, trim, concat, convert, scale, effects, audio ops, quality metrics, platform encoding, HLS, GIF, hardware acceleration, video analysis with sub-agents, and batch workflows.
Hello Retail integration. Manage data, records, and automate workflows. Use when the user wants to interact with Hello Retail data.
Implements DeepAgent components using LangChain's deepagents SDK. Use when building deep agents with create_deep_agent, configuring backends/subagents/skills/memory, need DeepAgent patterns (sandbox, HITL interrupts, long-term memory, subagent spawning), or ask "implement deep agent", "add subagent", "configure backend".
Fix Node.js CLI tools crashing inside Docker containers when host-installed node_modules require a newer Node version than the container provides. Use when: (1) "SyntaxError: Invalid regular expression flags" with /v flag in string-width or similar packages, (2) node_modules installed on host with Node 20+ but container has Node 18, (3) `npm install` with file: protocol creates symlinks that break inside Docker, (4) pnpm workspace packages become broken symlinks in containers. Covers version detection, Node binary mounting, and npm install strategies for cross-version compatibility.
Step-by-step guide for adding a new agent framework adapter to Agent Kernel. Use this skill when you need to integrate a new agent framework (beyond OpenAI, CrewAI, LangGraph, Google ADK). Covers creating the adapter module, implementing Agent/Runner/Module subclasses, adding optional dependencies, exports, and tests.
This skill MUST be invoked when the user says "brownfield integration", "extend existing code", or "modify existing file". SHOULD also invoke when encountering tasks with `[EXTEND]` or `[MODIFY]` markers, implementing against existing codebases, or integrating with established interfaces.
Use Scratch.jl for package-specific mutable data containers (caches, compiled objects, host-specific data). Use this skill when storing or managing scratch spaces in Julia packages.
Bootstrap a reusable agent-first Harness for a monorepo from a filled `HARNESS_PROFILE.md`. Use when a team has documented project purpose, repo map, workflows, boundaries, and quality priorities and wants Codex/Pi to generate the Harness structure, docs, plans system, and stable operating files around those facts.