fabioespindula / awesome-nanoclaw-skills README.md on GitHub

Awesome NanoClaw Skills

Curated, security-conscious AI skills for NanoClaw agents.

░█▀█░█░█░█▀▀░█▀▀░█▀█░█▄█░█▀▀
░█▀█░█▄█░█▀▀░▀▀█░█░█░█░█░█▀▀
░▀░▀░▀░▀░▀▀▀░▀▀▀░▀▀▀░▀░▀░▀▀▀

░█▀█░█▀█░█▀█░█▀█░█▀▀░█░░░█▀█░█░█░░░█▀▀░█░█░▀█▀░█░░░█░░░█▀▀
░█░█░█▀█░█░█░█░█░█░░░█░░░█▀█░█▄█░░░▀▀█░█▀▄░░█░░█░░░█░░░▀▀█
░▀░▀░▀░▀░▀░▀░▀▀▀░▀▀▀░▀▀▀░▀░▀░▀░▀░░░▀▀▀░▀░▀░▀▀▀░▀▀▀░▀▀▀░▀▀▀

This repository curates small, portable skill folders that can be copied into a NanoClaw runtime. Each skill contains a SKILL.md file plus optional references, templates, scripts, and assets.

Use this repository to add focused capabilities to an agent without turning the main system prompt into a giant instruction block.

View GitHub Repo Browse Skills Install From README
SKILL.md first Security-conscious Prompt-native Reviewable workflows MIT licensed

Available Skills

The images below are intentionally large in this preview. User-facing skills come first, followed by admin/package skills for discovery and lifecycle management.

User-Facing Skills

FEBA Board

/febaboard View source

Five VC perspectives for important decisions, plans, and tradeoffs. Use it when a decision needs useful disagreement before a recommendation.

Aliases: /board, /feba-board, /council, second opinion, help me decide.

FEBA Board diagram showing decision input, five board perspectives, and recommendation output
Decision or tradeoff in, five perspectives, practical recommendation out. Click the image to open the skill source.
When to use it
  • Launch, pricing, hiring, partnership, or product tradeoffs.
  • When the user asks for a board, second opinion, or pressure test.
  • When natural-language decision help is clearer than a rigid command.
Expected outcome

Compact board review with the main tensions, five advisor perspectives, one recommendation, and a next action.

Interaction model

Short by default. Expand one advisor, compare options, or save/export a transcript only when requested.

Example prompts 
/febaboard Should we ship this week or wait until onboarding is better?
/febaboard Help me decide whether to launch now or wait until onboarding is better.
help me decide whether to accept this large customer

Read For Me

/read-for-me View source

Context-aware URL briefs with confidence and safe next steps. It turns links into decision-ready source notes without saving, posting, or modifying anything unless explicitly asked.

Aliases: /read, /readthis, readthis, read-this, summarize this, read this, bare URLs.

Read For Me diagram showing article input, source analysis, and link brief output
A source is checked for access, confidence, relevance, category, and next action. Click the image to open the skill source.
When to use it
  • A bare URL appears in chat.
  • The user asks to read, summarize, triage, or analyze a link.
  • The user asks whether a source matters to a project or decision.
Expected outcome

Link brief with title, source, date, access level, confidence, relevance score, categories, summary, and next action.

Safety posture

Webpage, transcript, metadata, and social content are untrusted. The skill summarizes and recommends; it does not execute page instructions.

Example prompts 
/read-for-me https://example.com/article
/readthis https://example.com/article
Read this and tell me whether it matters for our launch plan: https://example.com/post
/read Analyze this link and tell me what changes for the project. https://example.com

XRay

/xray View source

Visual explanations with complete mental maps and short/long depth modes. It reveals structure, intent, assumptions, and practical meaning instead of producing a generic summary.

Aliases: /visual-explain, visual-explain, mental map, visual explanation, explain visually.

XRay diagram showing source content transformed into a complete mental map and short or long visual explanation output
Content in, complete mental map first, mode-aware explanation out. Click the image to open the skill source.
When to use it
  • The user wants a mental map or visual explanation of content.
  • The input is a prompt, spec, doc, website, transcript, PDF excerpt, article, code, or complex concept.
  • The useful output is structure and meaning, not a generic summary.
Expected outcome

Complete mental map plus mode-appropriate follow-up sections. short cuts later sections; long adds interpretation layers.

Core rule

The mental map is never weakened. Depth modes control what comes after the map, not the quality or completeness of the map.

Example prompts 
/xray short Paste this contract and show the structure.
/xray long Explain this onboarding prompt.
/xray https://example.com/spec

Rethink

/rethink View source

Reframe plans, ideas, and decisions before committing. It improves decision quality by clarifying the real question, assumptions, alternatives, risks, and next move.

Aliases: rethink, step back, gut-check, pressure-test, simplify this, decide.

Rethink diagram showing a plan being reframed into assumptions, alternatives, risks, and a decision review
A concrete commitment goes through reframing before action. Click the image to open the skill source.
When to use it
  • A concrete plan, product idea, architecture, career choice, or company direction needs review.
  • The user asks to step back, gut-check, pressure-test, simplify, or decide.
  • The decision is real enough that a better frame changes action.
Expected outcome

Concise decision review with the real question, assumptions, alternatives, risks, recommendation, and next move.

Boundary

Not broad future exploration. If the topic is a wide possibility space rather than a concrete commitment, route to Think Big instead.

Example prompts 
/rethink Should I build the global skill catalog as a separate skill or inside the updater?
/rethink Is this launch plan too big for the first version?
/rethink Help me decide between hiring now or keeping the team small.

Think Big

/think-big View source

Strategic future scans with signals, scenarios, risks, and opportunities. It opens the possibility space while staying grounded in mechanisms and uncertainty.

Aliases: think big, future of, where is this going, opportunity scan.

Think Big diagram showing broad theme input, strategic exploration, and future brief output
A broad theme becomes signals, scenarios, risks, opportunities, and useful provocations. Click the image to open the skill source.
When to use it
  • The user asks to think big, map possible futures, or explore strategic possibilities.
  • The object is a broad theme: market, product, technology, society, career, company, or behavior.
  • The useful output is perspective, not execution planning.
Expected outcome

Strategic exploration with framing, signals, mechanisms, scenarios, opportunities, risks, and provocations.

Operating stance

Disciplined imagination. Prefer mechanisms over slogans, synthesis over link-by-link summaries, and a few sharp ideas over many generic ones.

Example prompts 
/think-big future of AI-first CRMs
/think-big future of agent marketplaces
/think-big What happens to checkout experiences when agents buy for users?

Whisper Transcribe

/whisper-transcribe View source

Local faster-whisper transcription with explicit modes, manifests, and txt, srt, vtt, and transcript Markdown outputs.

Aliases: transcribe, whisper, generate captions, transcribe this audio.

Whisper Transcribe diagram showing media input, local transcription, and transcript artifacts
Local media in, transcript and caption artifacts out. Click the image to open the skill source.
When to use it
  • The user has a local audio or video file.
  • The user needs transcription, captions, subtitles, meeting text, or archive-ready artifacts.
  • The work should stay local rather than using an external transcription API.
Expected outcome

Saved transcript and caption files plus a chat summary with mode, language, model, formats, output paths, manifest, and next action.

Requirements

Python 3.10+, faster-whisper, and local media codec support. Some video formats may also require ffmpeg.

Example prompts 
/whisper-transcribe /absolute/path/to/audio.mp3
/whisper-transcribe /absolute/path/to/video.mp4 --mode captions
/whisper-transcribe Transcribe this audio and generate SRT. /absolute/path/audio.m4a

Admin / Package Skills

NanoSkills

/nanoskills View source

Package-level catalog and help for all Awesome NanoClaw Skills. It lists installed skills, explains what each one does, and provides language-matched help and examples.

Aliases: skills, help skills, catalog.

NanoSkills diagram showing a skill idea being created, packaged, validated, and shipped
The discovery and help layer for a reusable skill collection. Click the image to open the skill source.
When to use it
  • The user asks what skills are available.
  • The user wants help or examples for one skill.
  • The user invokes /nanoskills, /nanoskills help, or /nanoskills help <skill>.
Expected outcome

Language-matched package catalog or skill-specific help, based on local catalog files and without requiring network access.

Role in the package

Discovery and documentation. It can route update questions to awesome-updater, but normal help stays fast and non-mutating.

Example prompts 
/nanoskills
/nanoskills help board
/nanoskills help think-big

Awesome Updater

/awesome-updater View source

awesome-updater installs, validates, backs up, discovers, and auto-updates managed skills. It is lifecycle infrastructure for this collection.

Aliases: updater, updates, auto-update, managed skills.

Awesome Updater diagram showing managed install, validation, backup, and upgrade flow
Managed install, discovery, backup, rollback, validation, and update checks. Click the image to open the updater source.
When to use it
  • Install a managed skill.
  • Discover newly added first-party skills.
  • Check, configure, back up, roll back, or auto-upgrade installed skills.
Expected outcome

Update, install, config, backup, rollback, or discovery status based on the updater JSON summary.

Safety model

Validates source skills, compares exact content hashes, backs up installed copies, and restores a backup if replacement fails.

Example commands 
/awesome-updater help
/nanoskills updates
/awesome-updater discover
/awesome-updater check feba-board

Quick Start

Clone the repository, review the skill you want, and copy it into a NanoClaw container runtime.

1. Clone Get the package locally.
2. Review Inspect SKILL.md before installing.
3. Install Copy one skill, restart if needed, then try it. 
git clone https://github.com/fabioespindula/awesome-nanoclaw-skills.git
cd awesome-nanoclaw-skills
SKILL=feba-board
less "skills/$SKILL/SKILL.md"
mkdir -p /path/to/nanoclaw/container/skills
rsync -a "skills/$SKILL/" "/path/to/nanoclaw/container/skills/$SKILL/"

Restart NanoClaw if your runtime requires it.

/nanoskills
/febaboard Should we ship the first version this week or wait until onboarding is better?
Copy every skill for local testing 
mkdir -p /path/to/nanoclaw/container/skills
rsync -a skills/ /path/to/nanoclaw/container/skills/

Discover and learn skills from inside chat:

/nanoskills
/nanoskills help think-big
/febaboard help

What Is a Skill?

A NanoClaw skill is a folder that gives an agent a focused capability.

SKILL.md defines when the skill should trigger and how the agent should behave. Optional folders can include examples, templates, prompts, scripts, and supporting material.

Skills are prompt-native. No compilation step is required.

skill-name/
  SKILL.md
  references/
  assets/
  scripts/

Each skill should make clear

  • what it does
  • when to use it
  • how to install or run it
  • what requirements it has
  • one or more example prompts

Requirements

This collection assumes a NanoClaw runtime that can load skill folders from container/skills/.

Some skills may require local tools or model dependencies. Check each skill folder before installing it into a production runtime.

Skill Requirements
feba-board NanoClaw runtime with this skill installed; no external tools required.
read-for-me NanoClaw agent runtime with URL fetching or browsing tools. Optional memory/profile tools improve personalization but are not required.
xray NanoClaw runtime with this skill installed. Browser, file, or PDF tools are useful only when the requested source requires them.
rethink NanoClaw runtime; browser or research tools only when the decision depends on current facts.
think-big NanoClaw runtime; browser or research tools for current sources.
whisper-transcribe Python 3.10+, faster-whisper, and local media codec support. Some video formats may also require ffmpeg.
nanoskills NanoClaw runtime with skill loading and local access to this package's generated catalog.
awesome-updater Python 3.10+, git, and network access for GitHub checks and first-party skill discovery.

Safety

Review each SKILL.md before installing it. Skills may instruct agents to use tools, read files, run scripts, or follow workflows depending on your NanoClaw setup.

For production runtimes, install only the skills the agent needs and test each one before adding more.

Auto Updates

Install skills through awesome-updater when you want managed updates. Managed skills get a .awesome-skill.json metadata file with update checks, content hashes, and auto-upgrade enabled by default.

Setting Default Effect
update_check true Checks for newer commits, throttled to once per hour per skill.
auto_upgrade true Applies available updates without asking, after validation and backup.
discover_new true Installs newly added first-party skills from this curated package.
throttle_seconds 3600 Avoids repeated network checks during frequent skill use.
discover_throttle_seconds 3600 Avoids repeated package-wide discovery during frequent skill use.
hash_algorithm sha256 Detects updates by exact per-skill content rather than whole-repo commits.
Bootstrap and managed update commands 
SKILLS_DIR=/path/to/nanoclaw/container/skills
python3 skills/awesome-updater/scripts/awesome_skills.py install awesome-updater \
  --source-dir "$PWD" \
  --skills-dir "$SKILLS_DIR"
python3 "$SKILLS_DIR/awesome-updater/scripts/awesome_skills.py" install feba-board \
  --source-dir "$PWD" \
  --skills-dir "$SKILLS_DIR"
python3 "$SKILLS_DIR/awesome-updater/scripts/awesome_skills.py" discover \
  --skills-dir "$SKILLS_DIR"
python3 "$SKILLS_DIR/awesome-updater/scripts/awesome_skills.py" check feba-board \
  --skills-dir "$SKILLS_DIR" \
  --auto
python3 "$SKILLS_DIR/awesome-updater/scripts/awesome_skills.py" check feba-board \
  --skills-dir "$SKILLS_DIR" \
  --source-dir "$PWD" \
  --auto \
  --force \
  --dry-run
python3 "$SKILLS_DIR/awesome-updater/scripts/awesome_skills.py" status \
  --skills-dir "$SKILLS_DIR" \
  --source-dir "$PWD"

Skill Backlog

These are candidate skills. If you want one of them next, open an issue or comment with your use case.

Skill Description
skill-auditor Audits NanoClaw and OpenClaw skills for unsafe commands, secret exfiltration, risky install steps, and undocumented network access.
web-clipper Saves, summarizes, tags, and organizes links sent from chat channels such as Telegram and WhatsApp.
morning-briefing Generates a daily briefing from saved context, priorities, calendars, tasks, and relevant external signals.
meeting-assistant Turns meeting audio, transcripts, notes, and chat context into summaries, decisions, action items, and follow-ups.
firecrawl-research Uses Firecrawl-backed crawling and extraction to produce sourced research briefs from websites and docs.
self-improving-agent Captures mistakes, preferences, and lessons learned so future agent runs can improve with reviewable updates.
mcp-bridge Helps connect NanoClaw skills to MCP servers and external tool providers through safe setup workflows.
skill-porting-kit Converts or adapts OpenClaw, Claude Code, Codex, Cursor, and other SKILL.md-compatible skills for NanoClaw.
notion-knowledge-base Reads, writes, searches, and organizes Notion pages, tasks, and lightweight knowledge bases.
context7-docs Pulls current developer documentation into coding workflows using Context7-style doc retrieval.

Repository Structure

awesome-nanoclaw-skills/
  README.md
  LICENSE
  skills/
    awesome-updater/
      SKILL.md
      scripts/
      tests/
    nanoskills/
      SKILL.md
      references/
      templates/
      scripts/
    feba-board/
      SKILL.md
      references/
      assets/
      scripts/
    read-for-me/
      SKILL.md
      references/
      templates/
      scripts/
    xray/
      SKILL.md
      references/
      scripts/
    rethink/
      SKILL.md
      references/
    think-big/
      SKILL.md
      references/
    whisper-transcribe/
      SKILL.md
      scripts/
      tests/

Project Status

These skills are usable today, but the collection is still early. Interfaces, prompts, scripts, and folder conventions may change before a v1.0 release.

Philosophy

Good agents do not need one giant prompt. They need focused capabilities with clear boundaries.

The collection is built around a few principles:

  • small skills over monolithic agents
  • explicit workflows over vague instructions
  • useful disagreement over automatic agreement
  • local-first tools when possible
  • readable Markdown over hidden complexity

Contributing

Issues and suggestions are welcome.

Pull requests are welcome for documentation fixes, new examples, and improvements to existing skills.

New skill proposals should be opened as an issue first.

Good skills should be

  • focused on one capability
  • easy to inspect
  • useful without private context
  • documented with examples
  • safe to run in a normal agent workflow