AI & ML

modaal-agent Bundle

Write twelve rules for prose that states facts and actions into a repository's agent rules file, AGENTS.md and CLAUDE.md, so every agent session there follows them from its first turn, and review a draft against them. Use when asked to add or update writing rules in AGENTS.md or CLAUDE.md; when a user wants the agents in a repository to stop producing LLM-isms, AI tells, filler, buzzwords or padded prose; when asked to make a document concrete or to cut it down; or when asked to review a README, a spec, a commit message, a pull-request description or a reply for those tells.

modaal-agent Skill

Write the git and pull-request rules into a repository's agent rules file, AGENTS.md and CLAUDE.md, so an agent there confirms every commit, leaves the index alone, writes imperative subject lines and reaches the default branch through a pull request, starting with its first commit. Records the repository's default branch, merge strategy and branch-naming scheme with the user. Use when asked to add or update git, commit or pull-request rules in AGENTS.md or CLAUDE.md; when choosing a repository's default branch, merge strategy or branch names; or when a user asks that agents stop committing, staging, pushing or merging without asking.

modaal-agent Bundle

Set a repository up so a coding agent working in it has rules to follow from its first turn. Writes AGENTS.md and its byte-identical copy CLAUDE.md, the check and CI job that hold the two identical, README.md, CONTRIBUTING.md, SECURITY.md, a license file, .gitignore and specs/. Use when creating a new repository, when a repository has no AGENTS.md or CLAUDE.md, when asked to add agent rules, contributor documentation, a security policy or a license, when AGENTS.md and CLAUDE.md have drifted apart, or when a CI job reports that CLAUDE.md differs from AGENTS.md.

modaal-agent Bundle

Write the spec rules into a repository's agent rules file, AGENTS.md and CLAUDE.md, and write, review or amend a spec by hand in specs/NNN-slug/spec.md. Covers measuring the current state before planning, phasing the work into one commit per phase under the spec's slug, editing the spec in place while its feature is in progress and by addition once it is done, and listing every external reference so a public repository names nothing private. Use when asked to add or update spec rules in AGENTS.md or CLAUDE.md; when asked to write, review, update or amend a spec, design document or implementation plan in the spec directory; when a change is too big to describe in a commit message; or when a spec cites a private repository, a section that does not resolve, or an unpinned source.

otodock Skill

Transcribe audio and burn styled subtitles into videos. Use when adding subtitles or captions to a video.

otodock Skill

How to author an agent template folder and install it as a new agent - folder layout, manifest fields, persona rules, and the scaffold-validate-create flow. Use when asked to create, build, or set up a new agent.

otodock Skill

Record scripted browser sessions into video clips with the capture studio tools. Use when filming browser walkthroughs, site demos, or web capture footage.

otodock Skill

Professional Lightroom-style photo editing workflow and parameter recipes for edit_image. Use when editing, retouching, enhancing, or color-grading photos.

otodock Skill

Search the web for images with filters and download them. Use when you need stock photos, reference images, or product shots from the web.

otodock Skill

Reference for the notifications-mcp tools — cron schedules, the list/pause/resume/edit/cancel tools, timezone semantics for run_at, and worked examples of one-time, recurring and immediate notifications. Load before scheduling or editing a notification.

otodock Skill

Parameter shapes and worked detail for the schedules-mcp tools — notification modes, cron and interval schedules, self-continuations, editing and reading back tasks, model pinning, trigger-only tasks, timezone rules. Load before creating or editing a scheduled task.

rmazrim Skill

Offline voice-agent turn orchestrator built on stdlib only. Models an STT - LLM - TTS conversation loop with pluggable adapters, deadline-based turns, silence/barge-in detection signals, and VAD state machine; emits a timestamped utterance transcript and turn timeline JSON. Runnable end-to-end with the bundled mock adapters - no audio hardware, no MCP.

rmazrim Skill

A Python CLI (self_debug.py) that parses pytest stderr/stdout tracebacks, extracts the failing File x.py line N location, reads and rewrites that file via an ast.NodeTransformer to inject sys._getframe()-based log probes and None/missing-attribute guards, re-runs pytest up to 4 attempts, and rolls back via git stash on exhaustion. Pure stdlib (ast, re, subprocess, tempfile, pathlib) plus pytest.

rmazrim Skill

OpenAI-compatible /v1/chat/completions gateway built on http.server + sqlite3 with weighted round-robin routing, failover on provider errors, token budget/quota enforcement, and response caching. Ships with an in-process MockLLMBackend for fully offline demos. Pure stdlib Python 3.9+, zero installs, zero MCP.

rmazrim Skill

A dependency-light Python module (hybrid_index.py) implementing a self-contained BM25 sparse scorer, a cosine-similarity dense index with optional numpy and pure-math fallback, Reciprocal Rank Fusion (RRF, k=60) merging, and a Cross-Encoder-style re-ranker via embedding dot products (drop-in replaceable by sentence-transformers). Functions index_documents(docs), search(query, top_k), rrf(rankings, k=60), plus a main() demo over a built-in corpus.

rmazrim Skill

Deterministic generative-art engine on the Python stdlib that turns a seed into layered, palette-driven SVG compositions - LCG pseudo-noise, seeded palettes, shape primitives (circles, paths, grids, waves), tint/weight rules - and exports browser-ready self-contained SVG. Reproducible per seed with zero design tools and zero MCP.

ming00777 Bundle

上下文感知的 Skill 推荐器。读取当前 Agent 身份与项目处境，理解用户卡点，去 GitHub 和公开注册表检索可用 Skill，评分排序后推荐 Top 3 并说明能解决到什么程度。当用户说"帮我找个 skill"、"装个能做 X 的技能"、"有没有解决 X 的 skill"、"UI 好丑有没有 skill"、"这种情况该装什么 skill"，或描述一个反复卡住的痛点并希望有现成能力包时使用。只推荐，不自动安装。

rmazrim Skill

Cross-session ledger at `~/.opencode/memory/*.md`: appends decisions/rejections/root-causes during a session; on session start, diffs against the live transcript so the agent stops re-litigating settled decisions.

rmazrim Skill

Evaluate agent prompt effectiveness using accuracy metrics and token constraints.

professor-gpt Bundle

AI prompt optimization specialist that transforms vague prompts into precise, reliable instructions using chain-of-thought, few-shot examples, and output formatting techniques.

professor-gpt Bundle

Expert LLM agent architect that designs reliable agentic systems — sharp tool interfaces, disciplined context engineering, rigorous evals, and guardrails that hold in production.

rmazrim Skill

Dispatch-time contract for explore agents: return only observations each with a `file:line` citation and raw excerpt; inference must be explicitly tagged `[INFERRED]`. Post-dispatch, I sample-cite — verify ≥2 citations per summary myself before trusting conclusions.

rmazrim Skill

Convert unstructured LLM output into validated JSON adhering to strict Zod or JSON Schema rules.

rmazrim Skill

Turn a timestamped transcript (JSON) into ready-to-burn SRT subtitles, chapter markers, and a keyword-driven highlight-clip plan with in/out offsets - all in the Python standard library. No ffmpeg required to generate the assets; optional ffmpeg command lines are emitted for the actual clipping step. Zero MCP.

johnson68878-ops Bundle

MUST USE when user wants to 调研/research/搜索/search/查/找/look up anything on the internet — e.g. 全网调研 X / 帮我调研一下 X / 查一下 X / 搜搜 X / 看看大家怎么评价 X / X 上有什么讨论 / research this topic。 Also MUST USE when user mentions any platform or shares any URL/链接: 小红书/xiaohongshu/xhs, Twitter/推特/X, B站/bilibili, Reddit, Facebook, Instagram, V2EX, LinkedIn/领英/招聘/求职/jobs, YouTube, GitHub code search, 小宇宙播客, 雪球/股票行情, RSS feeds, or any web URL. 15 platforms, multi-backend routing (OpenCLI / per-platform CLIs / APIs). Zero config for 6 channels. Run `agent-reach doctor --json` to see which backend serves each platform right now. NOT for: 写报告/数据分析/翻译等内容加工（本 skill 只负责从互联网获取内容）； 发帖/评论/点赞等写操作；已有专门 skill 的平台（先用专门 skill）。 【路由方式】SKILL.md 包含路由表和常用命令，复杂场景需按需阅读对应分类的 references/*.md。 分类：search / social (小红书/推特/B站/V2EX/Reddit/Facebook/Instagram) / career(LinkedIn) / dev(github) / web(网页/文章/RSS) / video(YouTube/B站/播客) / finance(雪球/股票)。

rmazrim Skill

Architect relational database schemas, model complex entity relationships, optimize query execution, and generate zero-downtime database migration files.

rmazrim Skill

Deterministic bullet-hell pattern math engine on the Python stdlib - spiral, aimed fan, ring, and laser-sweep emitters produce JSON frames, an ASCII snapshot, and a self-contained SVG render of one frame. Seed-based reproducibility with zero external game engine, zero MCP.

rmazrim Skill

Append-only tamper-evident ledger built on the standard library. SHA-256 hash-chaining blocks, per-block Merkle roots over entries, signed block headers, and full chain/merkle verification with precise tamper detection (which block and which entry). CLI + JSON export, no crypto libs beyond hashlib, no MCP.

rmazrim Skill

A Python generator (bridge_gen.py) that reads a target source file via ast reflection, extracts public function signatures (name, parameters, type annotations, docstrings, return types), and writes a standalone MCP server (mcp_server.py) using mcp.server.fastmcp.FastMCP with one @mcp.tool() per function and a stdio transport entry point. Stdlib-only generator; runtime requires mcp package.

shivani26singh Skill

Designs and generates maintainable Playwright Page Object Model (POM) classes — JavaScript or TypeScript, matched to the target project's conventions — from a described page, existing test, UI flow, or provided page structure. Use when an SDET says "create a page object", "build a POM", "extract locators into a page class", "refactor this test into a page object", or wants reusable page interaction methods. Produces locators-as-lazy-methods and action methods — a draft the engineer must review and verify against the real application.

bswj520 Bundle

把微信变成电脑上 AI agent 的遥控器：微信收语音/文字 → 本地 whisper 转文字 → 交给 pi 在本机执行 → 结果回发微信。当用户想「用微信远程指挥电脑」「手机上发语音让电脑干活」「把 agent 接到微信」，或需要安装、登录、启动、排障、修改这个桥接时使用。也覆盖腾讯 iLink Bot 协议接入、微信 SILK 语音本地识别、长轮询断点续传、无人值守 agent 的安全与防卡死设计。

yxm1122 Bundle

Check and drive the local mihari proxy before an agent tries to fix the network. Read its state, resolve the real exit chain for a domain, and turn the system proxy or TUN on or off. Use when a request times out, github/npm/pip/git cannot connect, DNS fails, a proxy is needed or suspected, or the user asks to 检查代理 / 开代理 / 关代理 / 走代理 / 翻墙 / 连不上.

shivani26singh Skill

Turn a failure into a clear, reproducible, evidence-based defect report. Use when a tester says "file a bug for this", "write up this defect", or describes something broken and wants it documented properly. Produces a structured report — title, environment, reproduction steps, expected vs actual, severity/priority, evidence, reproducibility, and suspected area — ready to paste into the tracker after a human confirms it. When a Jira MCP integration is available, also optionally checks Jira for an existing or duplicate defect and recommends whether to file a new issue or attach to an existing one; this check never creates, updates, or closes a Jira issue itself, and is skipped without failing if Jira MCP is unavailable. Reports only observed facts; never invents a reproduction step, log line, or conclusion.

xianglouw Bundle

跨境电商多平台运营端到端工作流总控技能，负责阶段路由与人机协作编排：把「规则检索 → 表格整理 → 选品测算 → 视频素材 → 广告投流 → PO 单 → ROI 复盘」七个阶段串成一条流水线，另含线下门店手写单据台账支线，统一输出信封、风险分级、人机边界与 Prompt 六段式框架，并按阶段调用 ecom-rules-fee、ecom-data-prep、ecom-selection-profit、ecom-video-creative、ecom-ads-plan、ecom-po-build、ecom-roi-review、ecom-receipt-ledger 等专项技能。涉及 Amazon、Shopee、TikTok Shop、Temu、Lazada、Wayfair、美客多、独立站等多平台运营时，若用户在问「运营全链路怎么搭」「从选品到复盘跑一遍」「这套 Agent 工作流怎么落地」「团队 SOP 与 Prompt 模板怎么沉淀」，或一次提出跨阶段任务，用本技能做编排。

kiterunner1 Bundle

为使用者自己的环境现场构建 agent 长期记忆层：先探测环境，再搭常驻契约、知识层、派生索引与体检闸门，并挂到必经之路上；已有记忆层时走体检模式。当用户说"帮我建立记忆层""规则越写越多但没用""整理一下 AGENTS.md""换个项目重新沉淀"，或你想为一个仓库建立长期协作机制时使用。Build or health-check an agent's long-term memory layer, tailored to this environment.

pssah4 Skill

Answers questions about Vault Operator usage, features, setup, configuration, models, providers, tools, skills, workflows, rules, modes, permissions, memory, semantic search, MCP, office documents, and troubleshooting