local-first skills

1,560 tagged local-first, measured the same way as everything else here.

Browse within: LLM 297Multi-Agent 198agent-orchestration 196no-code 182coding-agents 179personal-ai 156discord-bot 155self-hosted 155slack-bot 155cli 153openclaw 126byok 114ai-design 113design-systems 113

extension-foundry

169

sdfdu/evopilot

Skill Claude CodeCodex

Analyze repeated EvoPilot observations and compile proven workflows into portable, versioned Agent Skills or other reviewed extensions. Use when improving or extending the agent itself.

not rated 2 changed 3d ago A 34 tokens original MIT

codex-channels

170

cafitac/codex-channels

Skill Claude CodeCodex

Skill "codex-channels" from cafitac/codex-channels, covering codex-channels, execution-first rule, fastest shortcuts, codex operator mode and fastest operator check.

not rated 2 4mo ago A 30 tokens original MIT

w48669664-bit/colla-agent

Skill Codex

Run Colla Agent, a traceable bounded collaboration loop in which the current Codex session implements a project and Antigravity plans and reviews it read-only. Use for complex or high-risk coding, design, refactoring, debugging, document, media, data, or multi-file project work when the user asks for Colla Agent…

not rated 2 yesterday A 130 tokens original MIT

pulse

173

linggen/skills

Skill Claude Code

GTM brain for solo founders launching products. Pulse reads the user's brief (identity, voice, hard rules) AND the configured workspace (README, /doc/, recent commits, source) to know the product as well as the user does. The page drives a three-step pipeline — Gather local (script), Gather web (agent + Fetch tools)…

not rated 2 changed today A 132 tokens original Apache-2.0

altai-cli

174

altaidevorg/altai-app

Skill Claude CodeCodex

Use the ALTAI terminal product to run agents, inspect durable workspace state, and automate safe coding workflows through human-readable or JSONL output.

not rated 2 today A 33 tokens original Apache-2.0

use-lanchu

175

lanchuske/lanchu

Skill Claude CodeCodex

Install and use Lanchu, a free, local-first control and trust layer that coordinates multiple AI agents over MCP.

not rated 2 16d ago A 29 tokens original MIT

vision-perceive

176

Yuhang-uestc/deepvision-local-mcp

Skill Codex

A multi-step local process for understanding images with a text-only AI model. It chooses a quick or detailed path and can combine image description, text reading, object finding, cropping, and checking results.

not rated 2 27d ago A 133 tokens original MIT

JNHFlow21/trove

Skill Claude CodeCodex

Fetch and enrich only exact cited media through bounded durable operations.

not rated 2 +1 changed 3d ago A 18 tokens original Apache-2.0

galdr-demo

178

Arakiss/galdr

Skill Claude CodeCodex

Summarize a git repository's recent changes and write them to a Markdown file. Use it when asked for a «change summary», «quick changelog», «what changed in the repo», or to dump git state to a file. Parameters: the repository and the output path.

not rated 2 yesterday A 61 tokens original MIT

cartograph

179

adder-factory/cartograph

Skill Claude CodeCodex

Explore a codebase through Cartograph v2's PostgreSQL code graph: intent-aware context, exact/BM25/hybrid search, callers/callees/impact, affected tests, live source, review, freshness, and bounded index administration.

not rated 2 changed today A 0 tokens original MIT

xiaojiou176-open/fileman

Skill Claude CodeCodex

Teach an agent to install Fileman's local MCP server, stay review-first, and use the safest manifest and batch-analysis tools before deeper mutation.

not rated 1 15d ago A 35 tokens original MIT

delx-memory

182

davidmosiah/delx-memory

Skill Claude CodeCodex

Local-first agent memory MCP. No API keys. Prefer MCP tools if connected; otherwise the package CLI. Use when the user wants Delx Memory data or actions through an agent.

not rated 1 yesterday A 40 tokens copy · 88% MIT

wellness-cgm

183

davidmosiah/wellness-cgm-mcp

Skill Claude CodeCodex

Local-first CGM MCP for agents. Prefer MCP tools if connected; otherwise the package CLI. Use when the user wants Wellness CGM data or actions through an agent.

not rated 1 10d ago A 39 tokens copy · 84% MIT

meet-live-assist

184

krystiangw/meet-live-assist-extension

Skill Claude CodeCodex

Live in-meeting assistant from a Google Meet transcript. Attach to the call over MCP and feed the user real-time help - answers to questions aimed at them, data/context, risks, talking points - using THIS agent's own domain context. Use when the user wants you to "watch my meeting", "help me live during the call", or…

not rated 1 21d ago A 0 tokens

contextdb-memory

185

mikhailbovt/ContextDB-Codex

Skill Claude Code

Use the local ContextDB MCP memory for durable project continuity: automatically recall and submit significant project-state candidates during substantial workspace tasks, organize an untrusted typed hierarchy, and handle explicit remember or recall requests.

not rated 1 14d ago A 44 tokens original Apache-2.0

corresync

186

nkiyohara/corresync

Skill Claude CodeCodex

Access guarded local multi-account mail, calendar, and tasks through Corresync MCP tools. Use whenever a request concerns Gmail, Google Calendar, Outlook, Microsoft 365, JMAP, IMAP, SMTP, CalDAV, an inbox or mailbox, email or messages, a calendar or schedule, availability, meetings, Teams links, reminders, to-dos…

not rated 1 today A 112 tokens original Apache-2.0

MyAgentHubs/aimemo

Skill Claude CodeCodex

Review GitHub PRs with learned code style preferences using aimemo persistent memory.

not rated 1 6mo ago A 23 tokens original MIT

onboard

188

ledq/resumery

Skill Claude Code needs its repo

Build the evidence bank from the user's raw career material (resumes, CVs, notes; pdf/md/txt/tex or pasted text). Ingests everything in bank/sources/ (or given paths / a paste), structures it into experiencebank.md + profile.json, interviews the user with a few high-value questions, and finishes with the evidence map.…

not rated 1 2mo ago A 88 tokens original MIT

consult

189

jgassens/codex-terminal-pro

Skill Claude CodeCodex

Ask other AI models (Claude Code and Kimi Code) for an independent opinion on a hard problem in this Home Assistant config - one, or several at once in parallel. Use when the user asks for a second opinion or another model's view, when two or more of your own attempts at a bug have failed, or before a change that is…

not rated 1 changed 3d ago A 114 tokens original MIT

codex-insights

190

Not-Dev-Tech/codex-insights-local

Skill Claude CodeCodex

Gere um relatório local e somente leitura sobre sessões, tokens, modelos, reasoning effort, profiles, MCPs, ferramentas, skills, prompts repetidos, finalidades e objetivos do Codex CLI. Use para auditoria de consumo e recomendações de configuração.

not rated 1 1mo ago A 57 tokens original MIT

remember

191

Utundry/sloplesscode

Skill Claude Code

A memory-saving skill that stores facts, preferences, decisions, tasks, and other context in Super Memory, a local semantic memory store.

not rated 1 1mo ago A 60 tokens original Apache-2.0

public

192

shawnhack/exocortex

Skill Claude CodeCodex

Skill "public" from shawnhack/exocortex, covering exocortex, tools, core, intelligence and knowledge graph.

not rated 1 1mo ago A 0 tokens original MIT

At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: