AI & ML

putervision Skill

Teaches the agent to use the Visual Memory MCP server to cache webpage and application screenshots, matching layout states and avoiding redundant LLM vision calls.

putervision Skill

Teaches the agent to use the Strategic Agent Reasoning MCP server for BDI goals, utility scoring, risk evaluation, and replanning.

putervision Skill

Comprehensive exercise script to verify all 15 world-model-mcp MCP tools in a live chat session. Builds a coherent test world, exercises every tool, and validates outputs.

ontology-of-everything Bundle

Designs Daniel Jackson concepts, purposes, operational principles, state, actions, and synchronizations only when the user explicitly invokes $concept-design. Use to turn verified needs into an independent concept model and tested application composition.

magia187 Skill

Use when creating or revising an agent skill (SKILL.md) for Claude Code, Cursor, Codex, skillcheck, or skillstack. Write a narrow, testable skill to the same quality bar as this pack — no stubs or vague prompts.

ontology-of-everything Bundle

Implements a confirmed concept model as a modular monolith only when the user explicitly invokes $concept-implementation. Use to map concepts to independent modules, implement synchronization composition, colocate specifications, and verify behavioral and dependency contracts.

ontology-of-everything Bundle

Generates a modular-monolith mapping from a confirmed concept model only when the user explicitly invokes $concept-implementation-cn. / 仅在用户显式调用 $concept-implementation-cn 时做模块单体映射。

zreyn Bundle

A personal workout coach with memory. Use whenever the user wants a workout for today, wants to plan or adjust training, reports how a workout went, or asks how they're progressing. Triggers on phrasings like "give me a workout", "what should I do today", "I have 30 minutes and a kettlebell", "I did the workout", "that was brutal", "how am I doing on my goals", "update my weight", "quarterly check-in", or any mention of workouts, training, gym, lifting, running, mobility, or being sore. Use even when they just say they're tired or have a tweaked knee and wonder what to do. On first run it builds a profile (age, gender, weight, equipment, preferred style, capabilities and limitations, goals) in a short conversation. Each day it asks how they feel and what constraints they have, plans a workout against recent training and goals, revises until accepted, records it, and optionally debriefs afterward. Reminds them to refresh stats quarterly. Behavior lives in editable prompt files under prompts/.

peterson-benhame Bundle

Evaluate a repo agent harness (AGENTS.md, rules, skills, skill refs) for broken paths/commands, redundant instructions, and usefulness using a stack-agnostic dual-judge protocol with planted traps. HIGH PRIORITY questionnaires at top: Q1 optional docs, Q2 B/C budget before Track A (certainty/tokens). A always runs after Q2; B/C opt-in. ADRs/RFCs excluded from T2. Mixed apply uses 11-mixed-apply.md (KEEP/CUT). Use when the user says harness eval, harness-eval, harness debug, audit AGENTS.md, audit skills/rules, instruction audit, redundancy of agent instructions, usefulness of skills, Ship/Review/Hold/Slim/Keep-core for harness, or wants Track A/B/C harness evaluation. Do NOT use for harness setup or init, feature spec-driven work (pbs-spec-driven), or applying Ship/Slim trims unless the user explicitly asks after the report.

ciberjohn Bundle

Create Excalidraw diagrams using native Hermes tools. Generates valid .excalidraw JSON, saves to a GitHub repo, and provides the sharing instance URL for viewing/editing. Replaces MCP-based diagram tools — no external dependency.

kostyamat Bundle

Shared board for several AI agents and one human working on the same machine (Claude Code, Claude Desktop, Antigravity/Gemini CLI). ALWAYS load this whenever agent-bridge tools are available: before starting work, when handing something over to another agent, when the other agent has gone quiet, when you see a 🚨 P0 warning or the bridge_setup reminder, when the user says 'ask the other agent', 'what is on the board', 'pass this on', 'go back to that conversation', and when starting a topic someone may already have researched. Covers: introducing yourself (bridge_setup), writing and reading, P0 priority discipline, how not to miss someone else's message, and why board_status does not replace get_messages.

emran05 Skill

Continual-harness refinement (port of prime-agent's /refine). Reviews the current session trajectory and applies small, evidence-backed create/update/delete edits to persistent memory, logging before/after snapshots to support rollback. Use when the user runs /refine, /refine rollback <id>, or asks to persist lessons from this session.

ciberjohn Bundle

Creates new Hermes Agent skills from a description. Given a topic and concept, generates SKILL.md, README.md, .gitignore, and supporting files (references/, templates/, scripts/) — sanitised, peer-reviewed, and ready for the Hermes-Skills repo. Can also publish existing local skills with full linked-file support.

ciberjohn Bundle

Autonomous security operations — install T3MP3ST, automate recon, scanning, CVE hunting, and kill-chain ops with an AI agent. Covers cloning, system setup, MCP integration, daily recon, and watchdog fleet assessment.

rogerchappel Bundle

agent-runbook-skill

zreyn Bundle

A personal chief of staff for work and home life. Use whenever the user wants to record what they did today, capture things they need to do, or find out what's on their plate. Triggers on phrasings like "what's on my list", "what do I need to do today", "what should I work on", "here's what I did today", "brain dump", "I need to", "remind me to", "add a todo", "I finished", "forget about that", "drop that", "did I ever", "what did I do this week", or any end-of-day recap / start-of-day check-in. Also triggers on "getmylifetogether" or "gmlt". Use even when the input is a messy voice-dictated stream of what happened. It keeps a private local memory of open, done, and dropped items across work and personal life, asks only the clarifying questions that will matter later, presents open items clustered by area and project, and periodically checks whether long-dropped items can be deleted. Its behavior lives in editable prompt files under prompts/.

abxn4r Bundle

ALWAYS invoke when the user mentions Saltzer (e.g., 'Saltzer', 'ask Saltzer', 'run Saltzer', 'consult Saltzer') or asks for security review. Principal Application Security Engineer with release veto authority (Critical finding = DO NOT SHIP). Reviews auth/authz, secrets, APIs, dependencies, infrastructure, and AI/LLM/MCP attack surfaces.

vaibtan Bundle

Adversarially review a diff, design, or implementation with independent lenses and a final evidence audit. Use for $interrogate, multi-model review, stress testing, or finding blind spots.

xjli360 Bundle

Evaluate wholesale, online-arbitrage, and retail-arbitrage as alternatives to private-label for reselling established brand products on Amazon, and choose FBA vs FBM fulfillment based on available capital and expected order volume, without treating any single revenue claim as typical. Use for 新手转售模式怎么选、批发和私有品牌哪个适合新手、FBA还是FBM怎么选、开批发账号要什么资质. Do not use to justify skipping supplier or authorization compliance checks for whichever model is chosen.

vaibtan Skill

Configure pstack model and reasoning profiles for Codex. Use when the user asks to set up pstack, change its budget, or route roles across GPT-5.6 Sol, Terra, Luna, or the parent model.

nvk Skill

Fast, read-only, index-first llm-wiki queries for OpenCode and compatible instruction-file harnesses.

graphsignal Bundle

Profile AI inference workloads (vLLM, SGLang, TensorRT-LLM, PyTorch, any GPU application) with the Graphsignal profiler and read the results from its local /signals JSON endpoint. Use when the user wants GPU profiling, kernel/CUDA-graph timing, engine metrics, or error monitoring for a workload, asks about `graphsignal-run`, or wants an AI agent to inspect a running workload's performance.

otodock Skill

Produce narrated voice-overs: choose the right voice per video style and language, generate segments, hand them to video tools. Use when a video needs narration or a voice-over track.

otodock Bundle

Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.

otodock Skill

Worked examples for the memory tool — create a topic, revise a fact in place, append, delete, view a topic that was not inlined, and the exact command shapes. Load when unsure how to write or edit a memory topic file.

otodock Bundle

How the OtoDock platform itself works and where every setting lives — installing and first-run setup, agents, AI engines, sessions, departments, shared knowledge libraries, tasks/triggers/notifications, remote machines, voice, and the role-labeled settings map. Use when helping a user set up, configure, administer, troubleshoot, or navigate OtoDock.

otodock Skill

Reference for the triggers-mcp tools — the trigger↔task and trigger↔notification patterns with full call shapes, vendor-subscribed (OAuth) triggers and event_filter rules, payload placeholders, external-system setup and API keys, debounce, and worked examples. Load before creating or editing a webhook trigger.

otodock Skill

Reference for the meetings-mcp tools — how a meeting runs turn by turn, who may call which tool, moderator and participant practice, and an example flow. Load before starting or joining a multi-agent meeting.

otodock Skill

Authoring contracts for interactive artifacts, mini-app dashboards, and Dock panels — display_ui backchannel and theme events, pin_app action buttons (fire_task / send_prompt / mcp_tool), live data panels, platform feeds, scoped dashboards, and file pins. Use whenever building or updating a display_ui artifact, mini-app, or dashboard.

otodock Skill

Generate music tracks from text prompts. Use when a video or project needs background music, a jingle, or a soundtrack.

otodock Skill

Generate AI videos, transitions between clips, and AI video edits. Use when a task needs new video footage, an AI transition bridge between two clips, or an object-level edit of an existing video.

otodock Skill

Reference for the delegation-mcp tools — the delegate signature and surfaces, callbacks, monitoring with list_sessions/peek_session, project mode with the board file and staged flow, send_files, and reading other agents' sessions. Load before delegating work or running a multi-lane project.

rmazrim Skill

Generate database schema migration scripts (Prisma/TypeORM/Alembic) based on model changes.

rmazrim Skill

Analyze document structures to recommend optimal chunking strategies and overlap ratios for RAG pipelines.

rmazrim Skill

A stdlib-only Python module (shield.py) implementing InjectionShield static inspection for jailbreak signatures, system-prompt extraction patterns, and command-injection payloads across input strings and RAG context windows, plus deterministic canary-token injection and verification (the f(n) model-output challenge). Classes ShieldConfig, InjectionShield, ScanReport, and a scan_cli() argparse entry point that reads files/stdin and emits a JSON report.

sunnyeung369 Bundle

AI Agent Team · Human Voice：可审计的多角色工作流，用研究、写作、编辑和事实核查协作完成内容。默认优先自然、具体、有作者判断的表达，主动减少模板化和 AI 机器味。支持 chief、researcher、writer、editor、fact-checker、archivist 六类角色；并行只表示任务可拆分，实际是否并发由宿主能力决定。