Showing 2941–2970 of 10169 skills
Use when working through architectural flaws documented in a paad/architecture-reviews/ report — selecting which flaws to fix, resuming a partial fix session across multiple sittings, or applying structural changes that need to be tracked back to a report. Not for producing that report — run /agentic-architecture first if there isn't one.
Ovid/paad
Use when the user asks which paad skills exist, what a paad skill does, which one fits their situation, or how to invoke one — including "what can paad do", "list the paad skills", "is there a paad skill for X", or a request for the arguments of a named paad skill
Ovid/paad
Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends Claude's capabilities with specialized knowledge, workflows, or tool integrations.
BintzGavin/helios
A coding agent on Windows: Git Bash path rewriting, PowerShell quoting and encoding, CRLF. Use when output comes back garbled, or it works here and breaks there.
phuryn/product-engineering
Before spawning a subagent or dispatching a build, review or research, and before trusting its report: the brief, the delegate, parallel runs, the handoff.
phuryn/product-engineering
Align IDE and agent tooling between Cursor, Devin, and Antigravity.
daledesilva/obsidian_ink
Sync Cursor, Devin, and Antigravity agent tooling from _reference_ide-setup into a target project.
daledesilva/obsidian_ink
Feature auto-development — clarify requirements, generate spec, dispatch dev sub-agent to implement and merge into current branch
ZhangShenao/harness9
Feature auto-development — clarify requirements, generate spec, dispatch dev sub-agent to implement and merge into current branch
ZhangShenao/harness9
给音频视频定制内容背景与标题,或要在视频上加一段动效图形时用:下横栏(姓名职务)、关键词字幕、大数字与图表、标注与引导、品牌贴片与片尾、倒计时与进度。不选模板,按规则写一个自包含的 HTML 画面(时间的纯函数),打包成代码包导入后放上时间线。只加静态文字、贴纸或已有的文字动画预设时不用;声波一律用内置的声波元素,不用本 skill 画;整条成片的规划看「从简报到成片」。
JimLiu/baocut
Only when the Agent SDK is asked for, an agent needs Claude Code's own tools, or the app already runs it: embed Claude agents in a Next.js app with the Agent SDK.
phuryn/product-engineering
This skill should be used when the user explicitly types "/end-session": the mandatory final step of any session started or resumed with /start-project or /resume-project, which the agent prints alone in its own block as the session's last line for the user to type, never invoking it directly. It updates the active project's PLAN.md (Status, kickoff, log), checks the PLAN against itself, and commits+pushes. Do not auto-invoke on vague phrases mid-session.
jellyrock/jellyrock
Execute a procedural implementation plan on the Sonnet model. Runs when the user types /sonnet with a plan path, or as a Sonnet sub-agent that a parent (usually /focus, right after it saves the plan) starts and supervises. Checks the plan is for this repo, walks its Approach literally, runs its Verification, commits, records the landing in the plan's project, and stops before push for an explicit OK. Not for judgment-heavy work: if the plan still has open forks, architectural decisions or "we need to figure out X" gaps, stay on the stronger model and finish the planning first. Not for investigation or recipe work that has its own skill (/ci-triage, /issue-triage, /new-setting and the like): the recipe is the better plan.
jellyrock/jellyrock
Create, track, and land convoys of related beads as primary work orders in the Gas Town multi-agent orchestration framework.
a5c-ai/babysitter
Systematic codebase exploration following the Iron Law - understand the problem before exploring code. Four phases with file-finder and web-researcher agents.
a5c-ai/babysitter
WASM-based instant code transforms for simple tasks, achieving 352x speedup over LLM inference with zero cost.
a5c-ai/babysitter
Multi-protocol consensus for agent swarms supporting Raft leader election, Byzantine fault tolerance, Gossip state propagation, and CRDT conflict-free merging.
a5c-ai/babysitter
Complexity-based task routing with Q-Learning optimization, Agent Booster WASM fast-path, and Mixture-of-Experts model selection.
a5c-ai/babysitter
Microsoft AutoGen multi-agent configuration for conversational AI systems
a5c-ai/babysitter
Chain-of-thought and step-by-step reasoning prompts for complex problem solving
a5c-ai/babysitter
Constitutional AI and safety guardrail prompts for aligned LLM behavior
a5c-ai/babysitter
CrewAI multi-agent orchestration setup for collaborative AI systems
a5c-ai/babysitter
LangChain ReAct agent implementation with tool binding for reasoning and action loops
a5c-ai/babysitter
LangChain tool creation and integration utilities for agent systems
a5c-ai/babysitter
LlamaIndex agent and query engine setup for RAG-powered agents
a5c-ai/babysitter
Token-efficient prompt compression techniques for cost optimization
a5c-ai/babysitter
Prompt injection detection and prevention for secure LLM applications
a5c-ai/babysitter
Apply to any non-trivial work, not just bulk work: edits, migrations, analyses, checks. Build the tool that does it or proves it (codemod, script, generator, or a skill your subagents follow) instead of working by hand. The tool is the artifact a reviewer can rerun.
neohaskell/NeoHaskell
Spawn three parallel review subagents over the active transcript, surface learnings, and route each to a concrete edit on an existing skill. Use when the user says reflect.
neohaskell/NeoHaskell
Keep authored or revised skill guidance durable with self-contained examples, owned operational dependencies, and separate run evidence. Apply to skill packages and their routing instructions.
neohaskell/NeoHaskell