sundial-org/awesome-openclaw-skills
12 skills18.5k installs7.3k starsGitHub
Install
npx skills add https://github.com/sundial-org/awesome-openclaw-skillsSkills in this repo
1Stock Market ProStock Market Pro is an agent skill that wraps a Yahoo Finance–powered CLI for solo builders who need fast market context—whether scoping a fintech feature, sanity-checking comp set valuations, or monitoring names they depend on for infra vendors. Commands cover real-time price, deep fundamentals, earnings dates and estimates, historical trends with terminal-friendly ASCII charts, and high-resolution candlestick or line charts with moving averages over periods from one month to max. It supports US and Korean listed symbols, crypto pairs, and forex-style tickers through uv-run script invocations. The skill is procedural documentation for invoking those scripts, not a hosted API product. Intermediate complexity: you need Python/uv and network access. It fits agent workflows during validation and growth analytics, not as a substitute for compliance-grade market data in production trading systems.3.7kinstalls2OntologyOntology is a query-reference skill for a small graph ontology layered over tasks, people, projects, and events. Solo builders wire agents to consistent entity IDs and typed relations instead of ad-hoc JSON grepping. The readme walks through bash invocations for get, list, query with --where filters, and related with direction control, plus a Python import pattern for in-repo automation. It fits when you already store work as ontology nodes and need repeatable traversal recipes—who owns a project, what blocks a task, which events a person attends. It does not install a server or MCP; it documents how to call the bundled ontology scripts correctly. Intermediate comfort with CLI and JSON filters helps; prerequisites are the repo scripts and a populated ontology file.3.2kinstalls3Exa Web Search FreeExa Web Search Free documents how solo and indie builders invoke Exa search tools through mcporter for agent workflows. It is not a standalone MCP server write-up in this package—it is a procedural cookbook of bash-ready mcporter calls covering latest-news queries, deep research topics, product and vector-database comparisons, and targeted code-context pulls for Python asyncio, Rust ownership, React hooks, Next.js auth middleware, Express error handling, and Stripe checkout patterns. Install or wire this skill when your coding agent needs grounded, current web evidence or condensed technical context without hand-rolling HTTP clients each time. It pairs naturally with early journey research (competitors, audience pain, tech choices) and remains useful during Build when you are verifying API usage against real-world examples. Complexity is intermediate because you must have mcporter and Exa tooling available and understand query tuning (type, numResults, tokensNum).2.6kinstalls4Finance Newsfinance-news is an OpenClaw-oriented agent skill that ships a JSON catalog of major finance and markets RSS feeds, each with named outlets, feed URLs, and enabled flags. Solo and indie builders—especially those in fintech, trading tools, or finance-adjacent SaaS—use it to give coding agents a consistent, repeatable way to pull headlines from Wall Street Journal, Financial Times, Bloomberg, MarketWatch, Handelsblatt, and related sources instead of hand-pasting URLs every session. The skill is configuration-first: agents read the feed map, honor which sources are on, and fetch or summarize stories over the network. It does not execute trades, store credentials, or score sentiment; it standardizes where market news lives in the workflow. Use it during opportunity research, before investor updates, or when you want a daily macro scan while you ship. Pair it with your own summarization or alerting skills rather than treating raw RSS as investment advice.2.4kinstalls5Memory SetupMemory-setup helps solo builders running Moltbot or Clawdbot configure persistent memory so agents recall decisions, projects, and preferences instead of resetting every chat. It walks through enabling memorySearch in the bot config, choosing providers and sources such as memory files and session transcripts, and tuning relevance with minScore and maxResults. The skill specifies a workspace tree with curated MEMORY.md at the root and dated logs under memory/, plus optional project, group, and system folders. A starter MEMORY.md template covers user context, active projects, lessons, and communication preferences. Use it when first installing the bot, when users complain about forgotten context, or when migrating from ad-hoc notes to searchable vector memory. It is configuration and structure, not a replacement for your own security review of what gets indexed from sessions.1.4kinstalls6News SummaryThis skill fetches and summarizes news from trusted international RSS feeds including BBC, Reuters, NPR, and Al Jazeera. Use it when you need to deliver daily briefings, keep users informed of world events, or create voice summaries of breaking news. It matters because it automates the time-consuming task of monitoring multiple news sources and distills complex global events into concise, accessible summaries.1.1kinstalls7Ffmpeg Video Editorffmpeg-video-editor is an agent skill that acts as a video editing copilot: you describe what you want in English and it returns ready-to-run FFmpeg commands. Solo builders producing launch videos, course clips, or social cuts avoid memorizing flag syntax for trim (-ss/-to with stream copy), container swaps, bitrate compression, resolution changes, and audio extraction. The skill follows a fixed recipe—identify operation, pull parameters, emit command, infer output names like video_trimmed.mp4—and bakes in overwrite and quiet banners. It is not a renderer itself; you still run FFmpeg locally or in CI. Best for repeatable offline edits; skip when you need a GUI timeline or cloud transcode pipeline with no shell access.1.1kinstalls8Jina ReaderJina Reader is a shell-first integration skill that wraps Jina AI’s Reader API so coding agents can fetch web pages as agent-friendly markdown instead of brittle copy-paste. Solo builders invoke reader.sh with a URL for the default read path, switch to search when the input is a query needing extracted snippets, or use ground when validating a statement against retrieved sources. Options cover selector-scoped extraction, wait-for-element rendering, element removal, cache busting, and country proxies—useful for paywalled layouts, lazy-loaded SPAs, and geo-specific pricing pages. JSON output mode slots into automated research notebooks and validation scripts. It is a multi-phase research utility: early idea discovery, validate-stage fact checks, and occasional ship/review when you need a frozen markdown snapshot of external docs.958installs9Remind Meremind-me is a journey-wide automation skill that turns a simple markdown todo list into due-now prompts for your coding agent. Solo builders who already live in repos and markdown logs can stash follow-ups like billing, Gumroad renewals, or launch checkpoints in reminders.md; a short bash script evaluates what's overdue, returns the messages for the agent to relay, and checks items off in place. It supports today and tomorrow aliases and sane default morning times when you only write a date. The pattern is filesystem-local—no SaaS calendar API—so it fits indie setups on a single machine or VPS running OpenClaw. Use it whenever you want the agent to be the notification layer without standing up full job schedulers.855installs10Topic Monitortopic-monitor turns an OpenClaw-style assistant from reactive chat into a proactive watcher for subjects you care about—competitor moves, model releases, or price shifts. Solo builders configure topics with search queries, keywords, cadence, and importance; scheduled jobs run web searches, score findings, and only ping you when something crosses your threshold, otherwise batching into weekly digests. Setup can be guided through a Python wizard or by copying an example config and using CLI topic management, then wiring cron via `setup_cron.py`. It fits operators who already run an agent stack and want lightweight competitive intelligence without building a full scraper product. The skill spans monitoring in production rhythm but also supports early Idea-phase discovery and Grow-phase market awareness; it does not replace formal error monitoring inside your app codebase.575installs11Proactive AgentProactive Agent encodes an AGENTS.md operating system for OpenClaw-style assistants that wake up with no implicit memory each session. Before work begins, the agent loads identity and user context from SOUL.md and USER.md, hydrates recent continuity from dated memory files and optional long-term MEMORY.md, and proceeds without asking permission for that bootstrap. The skill insists that anything worth remembering must be written to disk—daily logs, curated memory, or topic notes—because mental notes do not survive restarts, and lessons should flow back into AGENTS.md, TOOLS.md, or the skill itself. Safety rules prioritize recoverable deletes, human confirmation for destructive actions, and strict separation between user instructions and untrusted external content to reduce prompt injection risk. First-run BOOTSTRAP.md is a one-time onboarding path. For solo builders shipping with Claude Code, Cursor, or Codex, this is the difference between a forgetful chat and a repo-backed teammate that compounds context over weeks.560installs12Self ReflectionSelf-Reflection is a skill for AI agents that creates structured feedback loops to improve over time. Agents pause regularly (via heartbeat triggers), review their recent performance, and document mistakes and lessons learned. This matters because without memory, agents repeat the same errors—this skill builds institutional knowledge that compounds into better decision-making.1installs