AI & ML

polypoint Bundle

Use when driving or verifying things directly in PEP (POLYPOINT's Delphi/VCL personnel-planning Windows client) on a tenant box over RDP/SSH/WinRM — e.g. "verify yourself in PEP on ct-zinc-master", open the Stempeleditor, create a Zeitsumme/Block/Dienst, paint a shift onto the roster, rename a Dienst, read roster/GLAZ data, or change user permissions in UsrAdmin. Element-level automation (pywinauto over a loopback agent), not pixel clicking. Runs a preflight check and bootstraps the toolchain itself, then drives selector-based UI commands and flows. Requires the companion pep-driver toolchain.

polypoint Skill

Cross-cutting review step that runs after the focused steps. Connects the developer's already-accepted findings and surfaces concerns that fell between the dedicated steps (backwards-compat, operability, docs, config/migrations). Runs as a non-interactive sub-agent; the open-ended developer-directed part is handled by the orchestrator. Use as the last analysis step before drafting.

polypoint Skill

Review the tests in a pull request — useless/weak tests, missing coverage for changed behaviour, and whether integration tests are warranted. Runs as a non-interactive sub-agent producing a triageable findings report. Use as the test-focused judgment step of a PR review.

polypoint Skill

Review the core logic of a pull request against its ticket. Produces a plain-language summary, a ranked list of "review hot spots" (the most important logic, with clickable links to the code), a ticket-vs-implementation comparison, and triageable logic findings. Runs as a non-interactive sub-agent. Use as the first judgment step of a PR review.

postman-devrel Bundle

Find and rank developer influencers for an upcoming product launch. Searches across X/Twitter, YouTube, LinkedIn, Substack, Dev.to, and GitHub to identify technically credible voices in AI, APIs, agentic workflows, MCP, and developer tooling. Returns a scored and ranked list of influencer candidates.

postman-devrel Bundle

Weekly social media agent team — researches Postman blog, release notes, and trending news, creates 5 LinkedIn/Twitter posts, packages an employee advocacy kit, and auto-posts to Twitter on Mon/Wed at 10am PST.

byronwall Bundle

Compare multiple viable approaches before an agent makes a meaningful decision. Score the options from 0 to 10 only on dimensions that are specific and relevant to the task domain, then select the best option or design a stronger hybrid. Use when the user invokes `$consider-contrast`, asks to consider and contrast alternatives, wants to avoid tunnel vision, or wants a decision checkpoint inside a larger investigation, plan, design, or implementation task. Keep this as a proportional inline step unless the user asks for a full decision report.

byronwall Bundle

Refresh, search, review, and synthesize Byron's living v2c transcript archive without asking him to run commands or identify files. Use when Byron says "refresh transcripts", "load new transcripts", "process my voice memos", "run speech to text", "run the transcript LLM processing", "review the transcripts about...", "find the relevant ideas in...", "what did I say about...", asks whether a topic or project appears in his transcripts, or wants themes, decisions, requirements, opportunities, or changes of thinking recovered from voice memos.

dikamilo Bundle

Loads a dx shared reference document by topic. Invoked by other dx- skills to pull in shared reference material (change-md, effort-md, progress-format, plan-template, plan-data-model, plan-api-contracts, plan-failure-modes, interview, module-design, design-lenses, knowledge-layer, review-report, untrusted-content).

shaharsha Bundle

Use when work is finished but not yet committed and deserves an independent check — a plan or design doc about to be implemented, a feature or bugfix about to be merged, or a refactor touching money, auth, migrations, concurrency, permissions, or deletion. Also use when the user asks for a second opinion, a cross-model or adversarial review, "have Codex/GPT look at this", "review my plan", "check this before I commit", or says they don't trust the change yet. Requires the codex CLI.

shaharsha Bundle

Generate logos, icons, UI mockups, hero images, product shots, and other design assets using OpenAI gpt-image-2 or Google Gemini (Nano Banana 2 Flash, Nano Banana Pro). Use whenever the user asks to create, design, generate, mock up, render, or illustrate a visual asset, or asks for image prompt engineering. Picks the right model for the job, writes a model-specific prompt following the official prompting rules of each provider, calls the API, and saves the PNG to disk.

shaharsha Bundle

Use when you need to know what another agent session did, said or decided, get a message to one, start one from outside it, or when a session exists on disk but is missing from a client's list (the desktop app not showing sessions the VS Code extension shows, or the reverse). Also when a finding exists only in a session's prose and not in any diff, PR or ticket, or a .jsonl transcript is too large for grep or cat. Triggers include "what is that other session doing", "which session opened this PR", "my session disappeared", "did another session already fix this", "send a message to that session", "start a new session for me", "open this session in the desktop app", "why don't I see my sessions in the sidebar".

shaharsha Bundle

Use when authoring, editing, updating, trimming, or adding rules to a CLAUDE.md, AGENTS.md, CLAUDE.local.md, or other project-level instruction file for an AI coding agent — including small one-line changes. Triggers on "write a CLAUDE.md", "audit my CLAUDE.md", "trim my CLAUDE.md", "is my CLAUDE.md too long", "what should go in my CLAUDE.md", "why isn't Claude following my CLAUDE.md", "add this to CLAUDE.md", "edit CLAUDE.md", "update CLAUDE.md", "put a rule in CLAUDE.md", "remember this in CLAUDE.md", or the equivalent for AGENTS.md. Always load this skill before modifying any CLAUDE.md/AGENTS.md, not just when explicitly asked.

yansircc Skill

Minimum viable loop validation — before committing to a complex task, identify the core assumption and build the smallest possible falsifiable experiment. Throwaway, isolated, single-point penetration. Triggers: spike, minimum loop, preflight, tracer, verify first, feasibility check. Should also be proactively suggested when facing complex tasks with unverified core assumptions.

yansircc Skill

Adversarial debate — spawn an agent team with opposing positions to converge on truth through falsification

yansircc Skill

Negation mode — test whether a conclusion can survive counter-argumentation

yansircc Skill

Quickly pick up a project and understand the current development state

yansircc Skill

Compression mode — output the shortest generative program of the problem. Derive, don't retrieve.

yansircc Skill

Informed reset mode — redesign with known information to break anchoring bias

yansircc Skill

Analyze and optimize code following software engineering best practices

yansircc Skill

OpenAI Codex CLI command reference for building executable codex commands

yansircc Skill

Co-creation mode — AI uses structured questioning to help users sharpen vague ideas

yansircc Skill

Claude CLI command reference for building executable claude commands

yansircc Skill

Activate code hygiene mode — eliminate dead code and keep the codebase lean with every change

yansircc Skill

Less Is More: Addition by Subtraction

yansircc Bundle

AI agent foreman — orchestrate multi-step tasks and supervise agents using pawl. Supports Claude Code / Codex mixing. Triggers: orchestrate, supervise, foreman, start working, assign work.

yansircc Skill

Guide for packaging and publishing skills to a Claude Code plugin marketplace. Use when users say "publish skill", "publish to marketplace", "add skill to marketplace", "package skill as plugin", or "distribute skill".

yansircc Skill

Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends Claude's capabilities with specialized knowledge, workflows, or tool integrations.

yansircc Bundle

Delegate a bounded task to a local Claude-compatible CLI runtime through a stable adapter that returns normalized JSON plus durable artifacts.

yansircc Bundle

Record mistakes from bug fixes so agents never repeat them. Auto-validates on write. Triggers: learn from mistake, record mistake, remember this bug, don't repeat.

ysyecust Skill

审计活跃仓库、MCP服务器、插件、连接器、环境表面和工具设置，然后推荐最高价值的ECC原生技能、钩子、代理和操作员工作流。当用户希望帮助设置Claude Code或了解其环境中实际可用的功能时使用。

zhaji2333 Skill

当目标为 LLM 应用/Chatbot/智能客服/AI 助手/Copilot/Agent/RAG 知识库/多模态模型，或发现用户输入进入大模型提示、工具调用、知识库检索、对话记忆、文件解析，或需要测试提示词注入/越狱逃逸/System Prompt 泄露/训练数据与敏感信息泄露/RAG 检索污染/Agent 记忆污染/工具滥用与命令执行/SSRF/沙箱逃逸时调用。负责 OWASP LLM Top 10 (2025) 全域深度挖掘与对抗。

itspecialist111 Skill

dev/mcp-builder — skill payload

itspecialist111 Skill

Discover what capabilities exist in this org's skills registry — and the MCP bindings needed to call them.

kryptobaseddev Bundle

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction.

kryptobaseddev Bundle

Deep code optimization audit using parallel specialist agents. Each agent hunts for performance anti-patterns, inefficiencies, and suboptimal code using pattern-based detection (Grep/Glob) WITHOUT reading the full source code first — avoiding anchoring bias on existing implementations. Covers ALL optimization domains: database queries, memory leaks, algorithmic complexity, concurrency, bundle size, dead code, I/O & network, rendering/UI, data structures, error handling, caching, build config, security-performance, logging, and infrastructure. Use when asked to: 'optimize my code', 'find performance issues', 'audit code quality', 'speed up my app', 'find bottlenecks', 'code review for performance', 'find anti-patterns', 'improve code efficiency', 'reduce latency', 'optimize performance', 'code smell detection', 'find slow code', 'optimize this project', 'performance audit', 'code optimization'. Also triggers on: 'optimizar codigo', 'encontrar cuellos de botella', 'mejorar rendimiento'.