Write the plan before writing the code. Execute it in small batches. Review between batches. This is how you avoid rewriting the same feature three times.

Deep dives into a subtopic within an **existing** NotebookLM notebook using the NLM CLI (`/Users/user/.local/bin/nlm`). Decomposes the subtopic into sequential learning units and generates per-unit in

Writes Terraform code following strict module abstraction, file structure, and quality standards. Every module must represent a meaningful level of abstraction, use the latest stable releases, and pas

Perform an automated code review on recent changes. Checks for security issues, performance problems, error handling gaps, and code quality. Outputs a structured report with severity levels.

Systematically debug an error by tracing its origin through the codebase, identifying root causes, and suggesting targeted fixes. Produces a structured investigation log.

Generate personalized Upwork/freelance proposals from job postings. Paste a job description, and get a tailored cover letter that highlights relevant experience, addresses client pain points, and includes a concrete action plan. Optimized for high response rates.

Generate intelligent conventional commit messages from staged changes. Analyzes diffs, auto-detects scope and breaking changes, and commits with approval.

Run a comprehensive project health check covering dependencies, configuration, security, testing, and build integrity. Outputs a graded report card with actionable recommendations.

Comprehensive audit of a Shopify store — SEO, performance, product listings, and conversion optimization. Outputs an actionable report.

3D print models generated with Meshy AI. Handles printability analysis, slicer integration, multi-color printing guidance, and print-optimized download workflows. Use when the user mentions 3D printing, slicing, Bambu, OrcaSlicer, Prusa, Cura, Creality Print, or wants to print a figurine, miniature, or physical model.

Scans QuickBooks Online books via the CleanupOwl API to detect and fix bookkeeping issues. Use when the user mentions CleanupOwl, QuickBooks cleanup, QBO bookkeeping scan, fixing bookkeeping errors, running accounting scans, or applying solver fixes to flagged records.

hmrc-tax-vat

p6-earned-value

pine-strategy-builder

workflow-auditor

> **Compatible with Claude Code, Codex CLI, Cursor, Windsurf, and any SKILL.md-compatible agent.**

Pre-commit safety checks for AI-assisted development. Detects secrets, large files, merge conflict markers, sensitive files, and common AI-coding mistakes before they hit your repo.

Automated weekly self-audit of agent operations, cron health, memory gaps, and process failures. Use when performing periodic reviews, operational health checks, identifying automation gaps, surfacing wrong assumptions, or generating improvement recommendations. Triggers on: weekly review, meta-audit, operations audit, what broke this week, process review, improvement cycle, retrospective, week in review.

Use when analyzing large codebases (>200 lines), performing web searches for current info, or needing adversarial plan reviews.

A cybernetics × AI agent seminar skill. Grounded in *Cybernetics and Scientific Methodology* (Jin Guantao) and Wienerian cybernetics, it analyzes how cybernetic principles inform modern AI agent R&D, architecture review, and design guidance. Trigger when users mention cybernetics, feedback control, agent architecture, systems theory, black-box methods, negative feedback, positive feedback, agent design principles, steady-state systems, information entropy, purposive behavior, functional simulation, or Ashby's law of requisite variety. Also use when users want to review agent architectures from a cybernetic perspective, discuss feedback mechanism design, analyze agent stability and self-evolution, or organize cybernetics × AI agent seminars.

A double-loop learning execution review skill. Reviews the execution history after a task is completed and turns it into a concise retrospective, handoff, or improvement memo. Use this whenever a coding, writing, research, planning, debugging, or operational task has just finished; whenever files, commands, tests, or deliverables were produced; whenever the user asks to summarize what was done, review the execution, create a handoff, capture lessons learned, or prepare for the next iteration; and especially when the user needs not only a summary of what happened, but a judgment about whether the original goal, assumptions, success criteria, or working rules should change before the next pass. If there is execution evidence in the conversation, err on the side of using this skill instead of waiting for the user to ask explicitly.

Use when users want to commit changes, improve a commit message, split mixed changes into logical commits, clean up recent local history, remove accidentally committed files, squash or reword recent local commits, or generally sort out a local Git mess before review or a PR.

Pre-launch security audit for web apps. Runs OWASP checks via shell scripts, then AI analyzes findings. Covers rate limiting, hosting bypass, data exposure, headers, session security, AI risks. Complements /security-review.

Compare, evaluate, and upgrade SKILL.md skills. This skill should be used when introducing a new skill that overlaps with an existing one, checking if an installed skill has upstream updates, merging two skills with complementary strengths, evaluating whether a skill is worth installing, or auditing unused skills. Trigger phrases include 'compare skills', 'check for updates', 'merge skills', 'evaluate this skill', 'upgrade skill', 'audit skills'.

Comprehensive Seurat v5 (R) guide for single-cell RNA-seq and multimodal analysis. Covers installation, standard workflows (Normalize/SCTransform), clustering, integration (CCA/RPCA/Harmony), differential expression (FindMarkers/FindAllMarkers), visualization (DimPlot/FeaturePlot/VlnPlot/DoHeatmap), spatial transcriptomics (Visium/Visium HD/MERFISH/Slide-seq), CITE-seq, ATAC-seq, WNN, cell cycle regression, hashing/demultiplexing, sketch analysis, BPCells on-disk, pseudobulk, and format conversion. Use this skill whenever writing, debugging, or reviewing Seurat R code, building scRNA-seq pipelines, or looking up Seurat syntax, even for simple questions.

Run email A/B tests with statistical rigor. Use when testing subject lines, content variants, send times, CTAs, or measuring experiment significance.

Write emails people actually read. Use when crafting subject lines, structuring email content, writing CTAs, optimizing preview text, or improving open and click rates.

Understand what determines inbox vs spam placement. Use when emails land in spam, investigating Gmail Promotions tab, diagnosing deliverability issues, or optimizing engagement signals.

Diagnose email problems and plan email infrastructure. Use when emails land in spam, bounce rates spike, deliverability drops, open rates decline, providers throttle sending, you need to set up email for a new product, or you're not sure which email skill to use.

Use when choosing a memory architecture for an agent system. Decision tree covering context window, conversation buffer, vector store, structured DB, and episodic memory.

Use when understanding how agents discover and use MCP tools. How tool discovery works in the MCP protocol.

Use when writing prompts that generate better prompts. A meta-prompt that improves any AI conversation.

Use when choosing the right LLM for a specific task. Decision framework for model selection.

Use when deciding between on-device and cloud agent deployment. Cost and privacy tradeoffs.

Use when understanding the SOUL.md identity specification. The standard for constitutional agent identity.

'Author, fix, and review Eraser Diagram-as-Code DSL from natural language requirements using this repository specs. Use this skill whenever the user asks for Eraser DSL, diagram-as-code, flowcharts, ERD/schema diagrams, architecture/cloud diagrams, sequence diagrams, BPMN/swimlane diagrams, syntax fixes, icon/styling updates, or conversion from plain English process/system descriptions into runnable Eraser code.'

Embedded C development, motor control algorithms, Field Oriented Control (FOC), BLDC six-step commutation, sensing topologies, protection design, and STM32G4 hardware acceleration for digital drives. Use this skill when writing or reviewing firmware for PMSM, BLDC, and related motor-drive systems involving SVPWM, Park/Clarke transforms, current/speed/position cascades, sensorless observers (SMO/PLL), signal filtering, and hardware optimizations such as CORDIC and FMAC when justified by timing, determinism, or control-loop workload.

Role-based review system with lite mode for fast reviews and council mode for chaired deliberation. Reviewers speak from their own job function, can disagree with reasons, and the chair decides whether to approve, request changes, defer, or veto.

>

>

>

环境变量健康检查工具，扫描代码引用、检测硬编码密钥、验证格式

Multi-turn verification loop for LaTeX papers via external reasoning model. Each section agent conducts back-and-forth dialogue, pushing back on weak points, and produces structured reports.

Generate LaTeX documentation for mathematical research. Use when creating papers, notes, or documenting formulas, proofs, and verification results.

Complete guide for numerical mathematical investigations. Conducts structured interview to understand background, designs implementation and tests, creates detailed plan, then executes code, testing, and reporting.

Use this skill whenever helping someone build, design, or troubleshoot layouts in Elementor Pro V4 using the visual editor. Triggers on ANY mention of Elementor, page builder, Flexbox widget, Div block, section layout, widget configuration, responsive breakpoints, dark mode, scroll animations, Interactions tab, Variables Manager, Class Manager, Custom Code snippets, or any visual design work inside the Elementor panel. Also trigger when someone asks how to position elements, set spacing, apply typography, configure dark mode, prevent FOUC, or build scroll-triggered effects in Elementor. ALWAYS use this skill before suggesting any Elementor V4 implementation — it contains critical V4-specific panel structure, common AI mistakes to avoid, and production-tested patterns.

Review, create, update, and sync Markdown documentation so it stays accurate, current-state only, code-aligned, concise, and non-duplicative. Use when the user asks to write docs, update README/docs/*.md, sync docs with the codebase, review docs for drift, stale content, overlap, or inconsistent naming, or enforce documentation standards. Do not use for roadmap writing, speculative design, historical writeups, migration narratives, or changelog-style release narration unless the user explicitly asks for them.

Validate and format Markdown documentation structure. Use when Codex needs to check or fix heading numbering, prose wrap, table formatting, spacing, Markdown links, or other deterministic documentation formatting issues without doing broader documentation-content work.

DesignCortex is a perceptual frontend skill for Claude Code — exact CSS values, lookup tables, and formulas grounded in visual science so Claude stops writing averaged-out frontend code.

Provides SMS compliance, formatting, and delivery best practices for the Sendly API. Covers TCPA compliance, quiet hours, opt-out handling, E.164 phone formatting, message segmentation, and carrier verification. Applies when building SMS features that must be compliant and deliverable.