Progressive-disclosure instructions an agent loads on demand. Cheap until invoked, which is why the per-session figure on each card is usually two digits.
Use when starting a long or token-heavy task that could exhaust the Claude Code 5-hour or 7-day usage limit mid-run, or any long unattended run that must survive a usage-limit pause and pick the work back up afterward. Triggers on "otsukare", "watch my usage", "pause if I hit the limit".
Initialize or repair an agentMemory .memory directory and root instruction files for Claude Code, Codex, OpenCode, and Gemini CLI. Use when the user invokes /skills, asks to install agent memory, initialize .memory, set up AGENTS/CLAUDE/GEMINI memory pointers, or avoid repeating the agentMemory install command.
Audit and review all installed agent skills. Run on-demand or via cron to get a health report: skill inventory, usage tracking, version checks, dependency health, and actionable recommendations (keep, update, review, remove). Use when asked to review skills, check for skill updates, find unused skills, or audit the…
Huashu Design - Integrated design capabilities using HTML for high-fidelity prototypes, interactive demos, slideshows, animations, design variant exploration + design direction consultant + expert review. HTML is a tool not a medium, embody different experts (UX designer/animator/slideshow designer/prototyper)…
MLX Swift LM - Run LLMs and VLMs on Apple Silicon using MLX. Covers local inference, streaming, wired memory coordination, tool calling, LoRA fine-tuning, embeddings, and model porting.
Turn any content into a beautifully typeset, single-file HTML document in the user's established reading design (warm paper, serif body, blueprint accents), because the user finds raw markdown daunting to read. PRIMARY TRIGGER is the slash command "/paper [topic or content]": when the user types this, ALWAYS trigger…
A tool for editing large files, making batch changes, writing files from pasted or standard-input text, and checking, undoing, rolling back, or restoring edits.
A structured business-problem-solving method based on McKinsey-style analysis, including issue trees, hypotheses, and research reports or presentations.
First principles destructor — strips any statement, business model, product idea, process, or industry belief down to its physics/math/logic floor. Calculates the convention tax (what you pay because of habit vs necessity). Rebuilds from reality. Use when evaluating pricing, hiring timelines, product approaches, team…
A Chinese-language framework for judging decisions through practical constraints, evidence, returns, risks, and ordinary people's circumstances. It is associated with the direct analytical style of Hu Chenfeng.
Generates Figma Code Connect mappings for React or Angular repos, validates them with pre-checks and the Figma CLI, and can publish via figma connect. Use when a user provides a Figma file URL and a component repo and asks for Code Connect files, figma.config.json, or validation.
Use this skill when working in a project that uses Memory by Aictx as product-layer project memory. It guides the agent to query memory on demand mid-task, save product-meaningful changes after meaningful work, and sync memory at session end.
Integrate the Embinder WebMCP SDK into an existing product and make its browser experience AI-native. Use when a coding agent must inspect any web application or platform, inventory every page and user-facing function, expose real actions and live context through Embinder, add the resident agent, cursor, spotlight…
You are running an autonomous experiment loop that optimizes a system prompt for an AI agent evaluated on SimLab tasks. You will modify system-prompt.md, run the agent against tasks, measure the average reward, and iterate.
★not rated 24 3mo agoA0 tokens
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: