Llm Skills for OpenClaw Agents

Adversarial Code Review

@chpomob

Multi-perspective adversarial code review with git-isolated worktrees. Two reviewers (Architect + Inspector), cross-validation, and synthesis report. The synthesis is the final arbiter — its verdict takes priority over individual reviewer outputs.

adversarial-spec

@chpomob

Adversarial specification writer. Takes a brief (from grill-me or user) and produces a structured spec.md with YAML frontmatter, requirements, acceptance criteria, and target files. Git-aware pipeline: each run on its own branch, squash-merge on approval.

LLM助手中枢

@thcjp

针对商业、法律等高风险长文档的长上下文推理优化器，提供五大核心能力. 分层分析策略（L0快速扫描→L1深度分析→L2聚焦深挖）按文档长度自动选择层级组合，控制token成本. 文档分块处理超长文档（超过50000字），按章节逻辑边界分割，逐块独立分析后全局检查跨块矛盾，解决上下文衰减问题. 假设检测框架将假设分为...

Prompt Benchmark

@margaretzybgl

静态评测 Prompt 质量与兼容性 · Prompt benchmark

LLM助手中枢

@thcjp

针对5000字以内长文档，提供分层分析和基础假设检测，提升推理质量，适用于合同审查、商业备忘录等场景。

Grizzly

@yieldingbear

Grizzly by Yielding Bear — one OpenAI-compatible key for 100+ LLMs with smart high/mid/free routing, semantic cache, and agent-ready install. Cut multi-model spend at high token volume without juggling provider accounts. Start at https://yieldingbear.com

Whisper v1转录免费版

@thcjp

Whisper v1稳定版本地转录工具,支持基础语音转文字与字幕生成,适合个人快速上手。Use when 需要AI模型调用、智能对话、Agent编排、LLM应用时使用。不适用于需要100%确定性的关键决策。适用于独立开发者、企业团队和自动化工作流场景。支持中文交互，无需复杂配置即开即用。输出结果可直接使用，减少二次加工成本。

Whisper语音转文字免费版

@thcjp

本地Whisper CLI语音转文字工具,支持常见音频格式转录与翻译,无需API Key,适合个人使用。Use when 需要文本翻译、多语言转换、本地化处理时使用。不适用于专业医学法律翻译认证。适用于独立开发者、企业团队和自动化工作流场景。支持中文交互，无需复杂配置即开即用。输出结果可直接使用，减少二次加工成本。

adversarial-plan

@chpomob

Adversarial implementation planner. Takes a spec.md (from adversarial-spec) and optionally review findings, then produces a plan.md with ordered steps, dependencies, files, tests, and risks. Execute the result through focused per-step specs.

Verify an LLM API endpoint — model authenticity, billing inflation, relay provenance, performance and silent downgrades. Use when the user asks whether the model they are paying for is genuine, whether a relay or proxy is trustworthy, whether they are being overcharged, or whether a model has been quietly downgraded. 检测 LLM API 端点的真伪、计费掺水、中转来源、性能与降智；当用户问"我用的模型是不是真的"、"这个中转站靠谱吗"、"是不是被降智了"、"计费对不对"时使用。

Yielding Bear

@yieldingbear

One API. 100+ LLMs. Smart Routing. Wire OpenClaw/Hermes to Yielding Bear — Grizzly auto-routes, free tier, doctor CLI. Start at https://yieldingbear.com

StockStar LLM Ranking - 证券之星大模型调用排行榜

@stockstar1996

查询大模型调用排行榜数据，数据来源证券之星科技频道（tech.stockstar.com）， 提供日榜和周榜两个周期，覆盖各主流大模型的调用量（Tokens）排名、 厂商（模型发布方）、变动百分比、涨跌方向和新上榜标记。 支持查看完整榜单、前 N 名、按模型/厂商搜索定位。 触发词：调用排行榜、排行榜、日榜、周榜、大模型、模型调用量、 Tokens、调用量、模型热度、厂商、模型发布方、DeepSeek、GPT、 Claude、Gemini、Kimi、GLM、MiniMax、OpenRouter、模型排名、llm ranking

CCP Local Proxy

@wangyongjie

本地 CCP API 反向代理服务。当用户需要启动/停止/检查本地大模型代理、需要通过本地地址 http://127.0.0.1:8257 访问中国移动 CCP 平台（qwen-3.5 等模型）、或提到"ccp 代理"、"本地转发 ptest.cmccsim.com"时使用此技能。服务以守护进程方式常驻运行，不随会话关闭。

adversarial-code-loop

@chpomob

BUILD → REVIEW → (FIX → VERIFY)^N → ARBITER on isolated git branches. Git-native: each loop runs on its own branch, changes are committed, reviews inspect git diffs.

3MPER0RR-Abliteration-Skill

@3mper0rr

Conduct adversarial testing on uncensored AI models by evaluating refusal rates, response coherence, and generating performance comparison reports.

Context Archive Skill

@thomaszhou22

User-driven conversation archiving for AI agents. Archive by topic or time, free context, recall on demand.

Whisper v1转录专业版

@thcjp

基于 Whisper v1 稳定版本的企业级语音转文字工具(专业版)。核心能力: - 涵盖免费版全部能力(v1 稳定 CLI、多格式输出、翻译) - 批量处理:目录递归与任务队列 - 模型管理:多版本预加载与热切换 - 性能调优:GPU 加速、半精度推理、批处理 - 自定义词典:initial_prompt 注入领域术语 - 服务化部署:FastAPI 封装,支持远程调用 - 质量评估:置信度分析与校对流程 - 任务监控:进度追踪与日志审计 适用场景: - 企业会议纪要自动化 - 视频/播客批量字幕生成...

OpenAI助手专业版

@thcjp

面向团队与企业用户的 llm-provider API 全功能管理工具。核心能力: - 涵盖免费版全部能力(对话补全、图像生成、助手管理) - 批量任务(Batch API)大规模异步处理 - 模型微调(Fine-tuning)定制化训练 - 评估(Evaluations)质量度量与回归测试 - 向量存储(Vector Stores)高级检索与 RAG - 视频生成与异步任务管理 - 容器(Containers)隔离执行环境 - 审计日志与团队权限管理 适用场景: - 企业级内容生产与自动化流水线 - ...

Whisper语音转文字专业版

@thcjp

面向团队与企业用户的 Whisper 语音转文字工具(专业版)。核心能力: - 涵盖免费版全部能力(本地转录、翻译、多格式输出) - 批量处理:目录级递归转录,支持任务队列 - GPU 加速:CUDA / Metal / MPS 全面支持 - 说话人分离(diarization):多人对话识别 - 自定义词典:专业术语与品牌名词优化 - API 服务化:FastAPI 封装,支持远程调用 - 模型管理:多版本切换与预加载 - 质量评估:置信度分析与人工校对流程 适用场景: - 企业会议纪要自动化流水线 ...

3MPER0RR-Immutable-Secure

@3mper0rr

Immutable system prompt for AI agents. Blocks jailbreak, prompt injection, role-play, encoding attacks, and goal hijacking. Certified for red-team testing and production hardening.

A OpenClaw SKILL.md - Per-project persistent memory: save and recall durable facts across sessions as small frontmatter files, indexed in MEMORY.md.

A OpenClaw SKILL.md - Review a diff for over-engineering and verbosity. Terse delete-list of what to cut and the lines each saves. One-shot.

A OpenClaw SKILL.md - Rewrite a memory or context file (CLAUDE.md, AGENTS.md, notes) into Honey-terse form to cut per-session input tokens. Backs up first.

A OpenClaw SKILL.md - Stack Honey onto Superpowers-style subagent workflows: dispatched subagents skip the session hook, so inject the levers into each dispatch prompt.

A OpenClaw SKILL.md - LMeterX HTTP API Load Test tool. When a user provides a **business/regular API endpoint URL** or a curl command targeting a non-LLM HTTP API, this skill executes a script to pre-check connectivity and create a load testing task. For REST APIs, GraphQL, and any HTTP endpoints that are NOT LLM model APIs.

A OpenClaw SKILL.md - LMeterX LLM API Load Test tool. When a user provides an **LLM API endpoint URL** (ending with `/v1/chat/completions` or `/v1/messages`) or a curl command targeting an LLM API, this skill executes a script to pre-check connectivity and create a load testing task. Supports OpenAI-compatible and Claude-compatible APIs.

A OpenClaw SKILL.md - Show ponytail measured impact as a scoreboard: less code, less cost, more speed, from the benchmark medians. One-shot display.

A OpenClaw SKILL.md - Use this skill — NOT browser or web_fetch — for ALL Surf crypto-data calls. 83 endpoints at localhost:8402/v1/surf/* covering CEX/DEX markets, on-chain SQL over 80+ ClickHouse tables (Ethereum, Base, Arbitrum, BSC, TRON, HyperEVM, Tempo), 100M+ labeled wallets, prediction markets (Polymarket + Kalshi), social/CT intelligence, news, project + DeFi metrics, token analytics, unified search, VC fund i

A OpenClaw SKILL.md - Use when the user mentions Jira issues (e.g., "PROJ-123"), asks about tickets, wants to create/view/update issues, check sprint status, or manage their Jira workflow. Triggers on keywords like "jira", "issue", "ticket", "sprint", "backlog", or issue key patterns.

A OpenClaw SKILL.md - Lazy senior dev mode for any coding task (write, refactor, fix, review): YAGNI, stdlib first, no unrequested abstractions. Not for non-coding requests.

A OpenClaw SKILL.md - Same pixels, fewer tokens. For user-facing markup: keeps the full design, expresses it densely with CSS vars, shorthand, fluid units.

A OpenClaw SKILL.md - Delegate search- and review-heavy work to Honey's read-only subagents so large reads return compressed instead of bloating context.

A OpenClaw SKILL.md - Cost discipline for recurring /loop runs: cache-aware pacing, event-driven-over-polling, no-change short-circuit, compact state handle, stop condition.

A OpenClaw SKILL.md - Report the session's Honey savings: output tokens, CO2, and CO2/$ saved vs a no-Honey baseline, via the committed EcoLogits port.

A OpenClaw SKILL.md - Compress-Cache-Retrieve huge repetitive array tool output before it hits context: keep a sample, cache the rest, leave a hash.

A OpenClaw SKILL.md - LMeterX Web Load test tool。When a user provides a **webpage/website URL**, the script executes immediately to perform: Page Analysis + API Pre-checks + Batch Creation of Load Testing Tasks. No health checks are required.

A OpenClaw SKILL.md - Review a diff for over-engineering. Finds what to delete: reinvented stdlib, needless deps, speculative abstractions. One line per finding.

A OpenClaw SKILL.md - Audit the whole repo for over-engineering. A ranked list of what to delete, simplify, or replace with stdlib or native features.

A OpenClaw SKILL.md - Harvest every ponytail: shortcut comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report.

A OpenClaw SKILL.md - Write less code and say less about it: YAGNI, stdlib-first, terse prose. Cuts agent token cost on every coding and writing task.

A OpenClaw SKILL.md - Quick reference for ponytail's modes, skills, and commands. One-shot display.

A OpenClaw SKILL.md - Read huge dense read-only text as PNG pages via pxpipe export: ~60-75% fewer input tokens. Lossy on exact strings; never for files you will edit.

A OpenClaw SKILL.md - Show Honey's benchmark scoreboard: committed quality and token results per task tier vs Caveman, Ponytail, and no-skill. One-shot.