AI & ML

pantheon-org Bundle

Conduct a two-question session-end reflection to catch blind spots and under-investigated areas before concluding. The agent surfaces its lowest-confidence work items and identifies what the user might be missing, then offers to investigate. Based on a Reddit-post technique combining an LLM-suggested confidence audit with Sam Altman's blind-spot question. Do NOT trigger for brief single-answer queries, CI contexts, or in the middle of active work — only when a session appears to be concluding. Triggers: 'wrap up', 'we're done', 'conclude', 'session end', 'final review', 'before we go', 'sign off', 'that's all', 'anything else', 'finished', 'reflection', 'confidence check', 'blind spot', 'what are you missing', 'rate your confidence', 'review the session'.

marcelinero Bundle

Auditar sistemas de inteligencia artificial — gobernanza, ciclo de vida, datos, modelos, despliegue, monitoreo y cumplimiento conforme a ISO/IEC 42001, NIST AI RMF y EU AI Act. Activar siempre que se hable de auditoría de IA, AI audit, algorithmic audit, inteligencia artificial, machine learning, ML, deep learning, modelos predictivos, GenAI, LLM, sesgo algorítmico, fairness, explicabilidad, AI governance, MLOps, model risk management, ISO 42001, NIST AI RMF, EU AI Act, sistemas de alto riesgo, model cards, datasheets, validación de modelo, drift, hallucination.

oakoss Bundle

Create specialized agents for Claude Code. Use for agent, sub-agent, subagent, autonomous, task agent, custom agent, agent definition, agent config

mikiagent Bundle

Turn vague, underspecified, or high-level user requests into precise execution prompts by inferring safe context, asking only high-value follow-up questions, identifying missing constraints, assigning reference authority, defining invariants, and producing a concise structured brief with Goal, Context, Constraints, Workflow, and Done When. Use when a request is ambiguous, likely to produce inconsistent results, lacks acceptance criteria, mixes creative and technical intent, or would benefit from a stronger Codex/ChatGPT/image-generation prompt before execution.

mikiagent Bundle

Decompose complex coding, research, implementation, or verification work into a small number of focused child-agent tasks with separate context budgets. Use when independent evidence streams, subsystem ownership boundaries, parallel investigations, or Planner → Implementer → independent Verifier separation would improve reliability. Keep delegation shallow by default, give each child a self-contained objective/source scope/constraints/output contract/acceptance evidence, collect compact evidence packets instead of transcript dumps, and synthesize or verify results at the parent. Do not delegate tiny or tightly coupled work when coordination overhead exceeds the context benefit.

joseconti Bundle

Complete guide for creating a WordPress site from scratch using MCP Content Manager abilities. Use when the user asks to create a new WordPress site, build a website from scratch, configure WordPress for the first time, or set up an online store with WooCommerce. NOT for maintenance, migration, or updates on existing sites.

joseconti Bundle

Use when hardening WordPress sites, auditing security configurations, applying security measures, or working with the MCP Content Manager Security module: wp-config.php constants, .htaccess rules, HTTP headers, login protection, REST API restrictions, file permissions, and security scoring.

nvidia Skill

Prepare, create, publish, or edit a NeMo Relay pull request or its body using the repository template and contributor requirements. Do not use for code review or implementation without PR preparation.

nvidia Bundle

Use this skill when choosing or running NeMo Relay installation for the CLI, Python, Node.js, Rust, OpenClaw, or maintained framework integrations, or when explaining Hermes Agent's built-in Relay integration.

nvidia Skill

Change or review NeMo Relay GitHub Actions workflows where permissions, pinned actions, caching, reusable workflows, or release gates require repository-specific handling. Do not use for ordinary source changes that merely run in CI.

nvidia Skill

Author or edit NeMo Relay documentation or examples when repository-specific MDX, public API, integration, or release-history conventions matter. Do not use for review-only requests or incidental prose edits in code.

nvidia Bundle

Review NeMo Relay documentation, examples, or public text for NVIDIA technical writing style, terminology, and repository accuracy. Do not use for ordinary implementation review.

nvidia Skill

Add or change NeMo Relay packages, package metadata, module paths, generated artifacts, registry output, or release-facing build surfaces, including registering new unified-release surfaces with version automation. Do not use for a standalone project version bump, ordinary source builds, or tests.

nvidia Bundle

Compare NeMo Relay release refs and draft the current documentation release-notes page from verified repository evidence. Do not use for GitHub Release bodies or ordinary feature documentation.

nvidia Skill

Execute a requested NeMo Relay code freeze by cutting a release branch, updating nightly alpha configuration, bumping main, and preparing the required PR. Do not use for an ordinary version bump or release-note draft.

nvidia Bundle

Perform a NeMo Relay project release-version change across project-owned manifests, dependencies, lockfiles, and generated attribution surfaces while keeping just set-version coverage complete. Do not use for dependency-only updates or adding a package when no project version bump is requested.

nvidia Bundle

Use this skill when adding NeMo Relay typed wrappers, domain types, or provider codecs while preserving JSON middleware semantics and caller-visible behavior.

bex-co Skill

Autonomous maintenance routine that picks one tricky logic unit in a monorepo package — state machines, date/period math, sign conventions, dedup windows, pagination — models its states and edge cases exhaustively, proves any divergence from documented intent with a failing test, lands the minimal root-cause fix plus regression test, and ships via /ship. Use only when the user explicitly invokes /routine-logic-bugfixer or asks to hunt for logic bugs. SKIP for test-infrastructure failures and intermittent tests (routine-flaky-test-fixer) and for merely convoluted-but-correct code (routine-logic-simplifier). The core trigger is "model this tricky logic and fix what's provably wrong".

catch-the-wave Skill

Coordinate supervised Orca workers: threaded messages, blocking ask/reply, task dispatch, worker_done/escalation waits, task DAGs, decision gates, coordinator loops, and decomposing work across agents. Use `orca-cli` for full ownership handoffs — "hand off", "handoff", "handover", "give this to another agent", "another worktree" — unless asked to supervise, monitor, or coordinate a DAG, and for terminal control, lightweight terminal prompts, shell commands, Orca worktree management, and reading or waiting on terminals.

catch-the-wave Bundle

Design world-class iOS and macOS interfaces inspired by Teenage Engineering, Braun, and Buckminster Fuller. Orchestrates autonomous implementation via ios-design-implementer agent. Also handles "test X on simulator" requests via ios-test-planner and ios-test-runner agents.

catch-the-wave Skill

Scan codebase for LLM prompts vulnerable to injection attacks. Use when auditing security of prompts that accept user content, reviewing LLM integrations, or before deploying AI features.

mangowhoiscloud Skill

Reference when designing autonomous agent systems or deciding agent freedom/safety/context management/collaboration infrastructure. 10 design principles distilled from Karpathy autoresearch (autonomous ML experiment loop) + AgentHub (agent-native Git DAG). Triggered by "autoresearch", "agenthub", "ratchet" ("래칫"), "context budget", "dumb platform", "program.md", "overnight", "autonomous experiment" ("자율 실험"), "branchless", "single-file constraint" ("단일 파일 제약"), "fixed time budget" ("고정 시간 예산") keywords.

mmlong818 Skill

长时任务的 Harness 思维框架——当任务复杂、质量要求高、需要多轮迭代时，自动应用生成器-评估器分离、上下文重置、契约驱动、并行探索等模式。触发词：复杂任务、高质量输出、反复优化、迭代提升、多 agent 协作、长时任务、超越默认结果。

mmlong818 Bundle

Create, improve, and evaluate prompts for OpenAI image generation or editing, including exact text, reference preservation, transparent assets, diagrams, UI mockups, and iterative edits. Use when the user needs an image prompt or a prompt-driven image workflow; do not use for image understanding alone.

qoli Bundle

Use when Codex needs to ask DeepSeek v4 Pro a focused one-to-one question through a local Python wrapper instead of MCP. Default to DeepSeek for second-model consultation unless the user explicitly names Codex, codex exec, or the Codex provider; only then use the isolated Codex CLI backend.

vignesh-nagarajan-vn Bundle

Build and run eval harnesses for features that call a model, with a keyless mock lane in CI, a budget-capped live lane behind a key gate, thresholds in a versioned file, and judge scoring. Use when adding or changing evals, testing a prompt or model swap, gating a release on output quality, or investigating a quality regression. Runs scripts/check_thresholds.py.

vignesh-nagarajan-vn Bundle

Write, refresh, or audit a repo's CLAUDE.md onboarding contract so future sessions start from it instead of re-exploring. Use when a repo has no CLAUDE.md and will be worked in again, when asked to update or check it, when its paths or commands have gone stale, and after any change to layout, commands, or conventions. Runs scripts/verify_claude_md.py.

wzyn20051216 Skill

Run MATLAB R2026a tasks from Codex with reliable MCP-based execution, logs, artifacts, toolbox inventory, smoke tests, and reproducible validation. Use whenever MATLAB code, .m/.mlx scripts, MATLAB tests, local toolbox probing, or command-line automation is needed.

wzyn20051216 Skill

MATLAB R2026a workflow for signal processing, DSP, image processing, computer vision, lidar, medical imaging, deep learning, and AI-assisted reproducible experiments. Use for filters, spectrograms, wavelets, object detection, segmentation, classification, training, inference, and generated figures.

punyko8 Skill

分镜二次审查（双 Agent 协作）：主 agent 写分镜/提示词 → 审查 agent B 6 维对标剧本 → 自动迭代直到零遗漏。

punyko8 Skill

用于写即梦 Seedance 2.0 中文视频提示词。覆盖分镜提示词、参考图/视频/音频的角色分配、相机运动、动作节奏、声音设计、风格修饰,以及提示词优化。

punyko8 Bundle

Toffee Family 分镜头脚本双Agent闭环检验系统。用于 Toffee Family 儿童剧及同类低龄内容的剧本拆分、4–15秒视频生产段设计、对白逐字零删减与慢中快语速核验、场面调度、表演节拍、段内多镜切换、每段4–6格故事板、逐段AI视频提示词、黑白低完成度动画粗草绘制及量化质检，直至评分通过或强制终止。

punyko8 Bundle

第一性原理分镜双Agent闭环检验系统。基于大脑感知硬规律设计分镜，双Agent（写作者+质检者）闭环迭代，直至评分通过或强制终止。

punyko8 Skill

为 Toffee Family 儿童剧编写和优化 Seedance 2.0 / 即梦中文逐镜视频提示词。用于分镜转提示词、参考素材分配、机位与动作时序、动作对白同一时码、音效设计、代码块交付及逐镜自检；默认强制禁止任何 BGM、背景音乐、氛围音乐、配乐、旋律、节拍和乐器声，只生成对白、旁白与明确音效。

qredence Skill

Diagnose Fleet Pi chat runtime failures, including NDJSON streams, Pi sessions, Bedrock model errors, resource loader diagnostics, queueing, aborts, and Plan mode question flow.

qredence Bundle

Doc Gardening Skill