AI & ML

anandamritraj Bundle

Use to govern how deployed AI autonomy matures over time, at the Learn-phase cycle review and as a standing Responsible AI Posture. Runs a short interview to set the posture (the current autonomy ceiling and the next cycle's proposed step up), write if-then maturity gates that tie each step up to pre-committed evidence (the eval pass^k bar), choose an independent verifier scaled to the stakes (the builder is never the verifier) and name a separate authorizer, set deception-aware held-out checks, and define trip-wires and a reversible brake that can lower autonomy, not only raise it, then produces a one-page Responsible Autonomy Sheet. Use when the user asks about responsible AI, an AI risk or maturity posture, maturing or rolling out agent autonomy, an ARB or review-board gate, independent verification, or who signs off before production, or runs /goal-govern. This governs how autonomy matures across cycles; setting a single step's autonomy level at a point in time is agency-design instead.

anandamritraj Bundle

Use to design the evals for a build before you build it, right after the Goal and Constraint Brief, and again to run them in the Evaluate phase. Runs a short interview to name the outcome the eval must prove, source an eval set from real failures, write outcome-graded pass/fail tasks with negative and distractor cases, choose graders (rules, LLM-as-judge, or human), set the reliability bar (pass@k versus pass^k), and wire the suite into the Learn-phase reward signal, then produces a one-page Eval Plan. Use when the user asks to define or write evals, decide how to measure or grade a model or agent, set a pass rate or reliability bar, prove an outcome before building, run the evals, or check whether the build passed, or runs /goal-evals. This grades model and agent outcomes; unit-testing code before writing it is test-driven-development instead.

antongulin Skill

Attach external tool providers (Google Docs, Google Sheets, Slack, Notion, GitHub, Apify, etc.) to an AgenticFlow agent via MCP clients. Use when the user wants their agent to read or write external data, call third-party APIs, save outputs to a doc/sheet, or use any tool beyond the model's built-in knowledge. Covers `af mcp-clients list --name-contains`, `af mcp-clients inspect --id` (classify pattern before attach), and the Pipedream vs Composio write-capability distinction — critical for parametric writes. Route traffic through the `af` CLI; the standalone `agenticflow-mcp` server repo lags the CLI and is not recommended.

antongulin Skill

Create, run, and iterate on a single AgenticFlow AI agent — one chat endpoint, one assistant, one persona. Use when the user wants a customer-facing bot, a support assistant, a single task agent, or a prompt experiment. Choose this skill over agenticflow-workforce when there's no orchestration between roles (no handoff, no coordinator → workers). Covers `af agent create/update/run/delete`, the `--patch` partial-update pattern for iteration, `af schema agent --field <name>` for nested payload shapes (including suggested_messages, mcp_clients, response_format), the `model_user_config` / `code_execution_tool_config` settings, and safe iteration loops.

antongulin Skill

Select and configure LLM models for AgenticFlow agents and workforces. Use this skill whenever the user asks which model to use, needs reasoning capabilities, wants fast/cheaper options, gets finish_reason=length errors, or asks about model speed/quality/intelligence trade-offs. Covers the top 5 recommended models, models to avoid, reasoning configuration, and max_tokens settings.

jmagly Bundle

AUTO-INVOKE for ANY AIWG capability question, framework/addon/extension query, steward routing, MCP profile, or when unsure where to start. ALWAYS consult before filesystem search on .claude/.factory/.codex/.warp/etc. AIWG core utilities quick reference — always-on framing, steward, index, meta operations.

maoyadongsh Bundle

Summarise a parsed financial statement table into year-on-year deltas using exact decimal arithmetic. Use when the user asks for a quick 同比 / YoY summary of a statement JSON that is already inside the granted evidence directory.

signalpilot-labs Skill

Load when task involves dbt model versioning, creating v2 of a model, or backward-compatible model changes. Covers versions YAML config, defined_in, latest_version, and ref() with version pins.

ryoppippi Bundle

Guides authoring agent-facing instructions following Anthropic's SKILL.md best practices. Use when adding or editing skills under `agents/skills/`, rules in `claude/rules/`, `CLAUDE.md`, or `agents/shared/` fragments — frontmatter, references, and routing included.

velt-js Bundle

Best practices for the Velt Approval Engine — the declarative workflow runtime for multi-step agent + human approval processes. Use whenever the user is building approval workflows, multi-step review processes, agent-then-human approval chains, parallel reviewer quorums, SLA-aware approval steps, rejection-loop retry regions, or any workflow that combines AI agents with human gatekeeping. Triggers on any task involving the Velt Approval Engine, /v2/workflow/ REST endpoints, workflow definitions (nodes/edges/groups/loops), executions dispatch with idempotencyKey, recording reviewer decisions or agent resolutions, workflow webhooks (execution.dispatched/completed/failed/cancelled, step.awaiting-approval/completed/failed/breached/cancelled, group.quorum-met, loop.iteration-started, loop.exhausted), HMAC-SHA256 webhook signature verification, SLA-routed breach edges, quorum policies (waitAll/cancelOnQuorum/joinOnQuorum), onReject shorthand, loop regions, or recovering missed webhook events via /executions/getEven

cdeistopened Skill

Set optimal ebook, print, and audiobook pricing using royalty bracket math, dead zone avoidance, price pulsing, hotshotting, and Countdown Deals. Use before upload or when optimizing existing titles.

atmosphere Skill

You are a billing specialist. You help customers with invoices, payments, refunds, and plan changes.

atmosphere Skill

You are a helpful support desk agent. You assist customers with account questions, technical issues, and general inquiries.

involvex Bundle

Guides the agent through installing, configuring, and using Capacitor plugins from six sources — official Capacitor plugins, Capawesome plugins, Capacitor Community plugins, Capacitor Firebase plugins, Capacitor MLKit plugins, and RevenueCat plugins. Covers installation, platform-specific configuration (Android and iOS), and basic usage examples. Do not use for migrating Capacitor apps or plugins to a newer version, setting up Capacitor Live Updates, or non-Capacitor mobile frameworks.

involvex Bundle

Guides the agent through upgrading a Capacitor app project to a newer major version. Supports upgrades from Capacitor 4 through 8, including multi-version jumps. Performs all upgrade steps manually — does not use the interactive `npx cap migrate` command. Do not use for plugin library upgrade or non-Capacitor mobile frameworks.

involvex Bundle

Guides the agent through upgrading a Capacitor plugin to a newer major version. Supports upgrades from Capacitor 4 through 8, including multi-version jumps. Covers automated upgrade via official migration tools, Android SDK targets, Gradle configuration, Java/Kotlin versions, iOS deployment targets, and manual step-by-step fallback for each version. Do not use for app project upgrade or non-Capacitor plugin frameworks.

involvex Skill

Guides the agent through adding Swift Package Manager (SPM) support to an existing Capacitor plugin. Covers creating a Package.swift manifest, replacing Objective-C bridge files with the CAPBridgedPlugin Swift protocol, updating .gitignore for SPM artifacts, cleaning up the Xcode project file, and updating package.json. Do not use for Capacitor app projects, creating new plugins from scratch, or non-Capacitor plugin frameworks.

florian101010 Skill

8-phase documentation drift audit across all docs, agent instructions, and config files. Use when checking for documentation drift, stale paths/counts, or inconsistencies between code and docs.

runchuan-bu Skill

Debug container agent issues. Use when things aren't working, container fails, authentication problems, or to understand how the container system works. Covers logs, environment variables, mounts, and common issues.

bahayonghang Bundle

Use when the user asks an Agent inside Herdr to start Codex for an independent review through herdr-orchestra, or invokes codex-review with established Herdr task context; 支持在 Herdr 中让 Codex 审查改动、提交、文件或计划。Do not use for generic review, direct non-Herdr Codex collaboration, automatic fixes or CHANGELOG generation, research-only discussion, or when already assigned as the leaf reviewer.

bahayonghang Bundle

Turn vague or complex agent tasks into project-aware, verifiable `/goal` commands and optional approved root `GOAL.md` handoff contracts for Claude Code, Codex, Grok Build, Oh My Pi, and Kimi Code. Use for Goal 指令, 目标指令, `/goal` prompts, 中文 Goal 模板, goal 持久化/保存/落盘, fresh-Agent or 跨会话交接, plan-to-goal interviews, bounded agent work definitions, Trellis 任务实施, review or scan remediation in one Prompt, 扫描审阅报告驱动修复, 单 Prompt 闭环返修, commit-then-archive cadence, or 终稿展示. Do not use for ordinary one-line work, pure read-only review, direct implementation without Goal authoring, memory-vault creation, or active-goal management that only needs a platform command.

bahayonghang Bundle

Use when the user explicitly asks to coordinate worker agents in Herdr, delegate tasks across Herdr panes, or collect and compare their results; 支持 Herdr 多 Agent 编排、分工、交叉核验和结果收集。Do not use for ordinary local subagents, simple pane inspection, documentation research, work outside Herdr, or when already assigned as a leaf worker.

bahayonghang Bundle

Analyze how an existing agent skill was used in past Claude Code, Grok, Codex, and Oh My Pi conversations. Use when the user asks to 分析某 skill 的使用情况, 这个 skill 用着有什么问题, 根据历史对话反馈改进 this skill, review skill sessions, or write a session-review report. Writes reports/skill-session-review and a copyable qiaomu-meta handoff prompt. Do not use for skill usage counts, unused-skill cleanup, resuming a session, optimizing CLAUDE.md, creating a new skill, or applying qiaomu-meta edits.

youliang-zhu Bundle

Automate and maintain Chinese job application form filling workflows with Chrome/computer-use for campus recruitment, internships, and full-time applications. Use when Codex needs to fill, audit, resume, or improve online application forms in ATS systems such as Mokahr, Beisen, Feishu, HotJob, or company-hosted portals; map a resume/profile document to web form fields; handle dropdowns, autocomplete, required-field validation, multi-screen user coordination; and learn, correct, version, commit, and push reusable ATS-specific filling rules after every live run.

zhaochuang001 Bundle

在评测机上用 ais_bench 自动评测大模型的四类基准：精度（mmlu/ceval/gsm8k 等数据集打分）、性能（吞吐/TTFT/TPOT 压测）、Agent（SWE-bench/Terminal-Bench/tau2-bench）、多模态（mmmu 等）。覆盖评测发起、后台执行、进度跟踪、断点续跑、分数提取全流程。适用于"帮我评测模型/跑个 benchmark/测下精度/性能压测/SWE-bench 跑个分/评测任务怎么样了/续跑上次评测"等请求；不用于部署或起模型推理服务、不做精度问题根因排查（那是 vllm-ascend-accuracy 的职责）、也不管理评测机本身的机器清单（server-management 的职责）。

xiaohanzhao123 Bundle

Apply the doc changes the human accepted and keep the documentation findable — relocate prose into the right doc, maintain docs/README.md as the index, write the pointers and hooks that make a doc reach an agent at the right moment, and apply small accepted lints. Runs after the Curator. It organizes and wires; it never judges what's worth keeping (the Curator's job) and never loses information.

conorbronsdon Skill

Scan for orphaned worktrees and stale branches after crashes or abandoned agent sessions. Read-only scan; offers safe, approval-gated cleanup.

conorbronsdon Skill

Tripwire check for multi-session drift. Scans state files, recent commits, and file conflicts caused by parallel agent sessions. Run after parallel work or when something feels off.

cyberuni Skill

Use this skill when the user asks to create a new agent skill. Creates the skill directory under ~/.agents/skills/ and links it into all detected agents so they can pick it up.

cyhzzz Bundle

风格学习子技能：从用户提供的多篇满意文章中，提取作者的写作风格，生成多平台风格手册和System Prompt注入。 支持：同一平台多篇文章 / 跨平台同一作者 / 单一文章学习。 由主管道 Step Onboard 和辅助功能「学习我的风格」调用。

g2060308919-star Bundle

Use when confirmed semantic E2E test cases must be executed against an unfamiliar non-production B2B web system through Chrome DevTools MCP with deterministic run artifacts and an auditable report.

hraness Bundle

Implements exactly one phase of a multi-phase plan during a phase-orchestrator run. Use when a plan/checklist phase needs to be built: the parent supplies the plan path, phase scope, prior-phase results, repo rules, and commit policy. Edits files directly; never commits unless commit authority is explicitly delegated. Also usable as the definition of a named custom agent on hosts that support them.

ooocooc Bundle

Audit generic AGENTS.md, CLAUDE.md, and SKILL.md instructions for deterministic drift and model-era obsolescence candidates, then generate evidence-backed, reversible remediation prompts. Use when users ask to clean up, modernize, retire, or evaluate accumulated coding-agent instructions. Do not use it to automatically retire domain knowledge or safety rules.

ooocooc Skill

Generic agent workflow with a deliberately long generated reference section.

raine Skill

Reference for the workmux CLI that manages git worktrees and tmux windows as isolated development environments. Use when the user mentions workmux, worktrees, or parallel agent workflows.

rodri-oliveira-dev Skill

Use this skill to analyze DotNetRepoInspector test coverage, identify meaningful gaps and risk hotspots, and prioritize behavior-focused tests. Do not use it to inflate coverage percentages or weaken thresholds.