The fog clears when you stop repeating yourself.

Remember the first time you realized you'd been copy-pasting the same prompt for the third week in a row?

You thought: "I should turn this into a skill."

You didn't.

You typed it again the next day. And the day after that. The intention was there — the action never followed.

skill-fog closes that gap. It watches what you ask Claude to do, counts the repetitions, and when you've typed the same thing enough times, it asks: "Want me to build a tool for this?" Then it actually builds it.

Demo

[demo.gif — coming soon]

Want to see it in action? Star the repo and we'll prioritize the recording.

How it works

  Your Claude Code session
           │
           ▼
  ┌─────────────────┐
  │   Stop Hook     │  ← Fires silently at every session end
  │  (stop.sh)      │
  └────────┬────────┘
           │
           ▼
  ┌─────────────────────────────┐
  │  Normalize + Hash           │
  │  "fix error on line 23"     │  → "fix error on line NUM"
  │  "fix error on line 145"    │  → same pattern ID
  └────────┬────────────────────┘
           │
     ┌─────┴──────┐
     │            │
   < 3x        >= 3x, 2+ sessions
     │            │
  Keep        ┌───▼────────────────┐
 watching     │  pending/{id}.json │
              └───────┬────────────┘
                      │
              Next session start
                      │
                      ▼
      ┌───────────────────────────────┐
      │  "I noticed you've asked me   │
      │   to fix line errors 5 times. │
      │   Want me to build a skill?"  │
      └───────────────────────────────┘
                      │
           ┌──────────┼──────────┐
           ▼          ▼          ▼
        skill      command     agent
           │          │          │
           └──────────┴──────────┘
                      │
                      ▼
             ~/.claude/ (done.)

Installation

# npm (recommended)
npm install -g skill-fog

# curl (no node required)
curl -fsSL https://raw.githubusercontent.com/pabang0620/skill-fog/main/install.sh | bash

Requirements: bash >= 4.0, jq (or python3 as fallback), Claude Code

How it grows you

Step 1 — Detect

Every time your Claude Code session ends, a Stop hook silently reads your message history. It normalizes the text (strips filenames, numbers, specifics) and hashes the pattern. No data leaves your machine.

Step 2 — Propose

When the same normalized pattern appears 3+ times across 2+ sessions, skill-fog promotes it to pending. The next time you open Claude Code, it surfaces the suggestion naturally — no interruptions, no popups.

Step 3 — Generate

Say yes. skill-fog writes the actual SKILL.md, slash command, or agent definition into ~/.claude/ — ready to use in your next session.

CLI Commands

| Command | Description | |---|---| | skill-fog status | Show all tracked patterns and their counts | | skill-fog review | Browse pending patterns and decide what to build | | skill-fog list | See everything skill-fog has generated for you | | skill-fog clean | Remove old, rejected, or stale patterns | | skill-fog doctor | Diagnose installation and hook registration |

Why skill-fog?

"Hermes remembers. OpenClaw automates. skill-fog evolves."

| | Hermes | OpenClaw | skill-fog | |---|---|---|---| | What it does | Persistent memory across sessions | Automated code actions | Detects your patterns, builds your tools | | Requires server | Yes | Yes | No — pure bash | | Learns from you | No | No | Yes — gets smarter every session | | Output | Context | Actions | Reusable skills, commands, agents | | Runtime | Custom | Custom | bash + SKILL.md — Claude Code native | | Setup complexity | Medium | Medium | One curl command |

skill-fog doesn't try to be a memory system or an automation layer. It's a pattern → tool compiler. The more you use Claude Code, the more skill-fog learns what deserves to be a permanent tool in your workflow.

What people are saying

"I had the same code review prompt copy-pasted in my notes for two months. skill-fog turned it into a slash command in 30 seconds." — Early tester

"The threshold system is exactly right. 3 times across 2 sessions means it's actually a habit, not a coincidence." — Beta user

Privacy

All pattern data lives in ~/.skill-fog/ on your machine. Nothing is sent anywhere.

• API keys, tokens, emails, and secrets are masked automatically before storage
• You can inspect (and delete) everything with skill-fog clean
• Uninstalling removes all data

Roadmap

• [x] Stop hook pattern collection
• [x] Threshold-based pending promotion (3x / 2 sessions)
• [x] Skill / command / agent generation
• [x] CLI (status, review, list, clean, doctor)
• [ ] Interactive review TUI (skill-fog review --interactive)
• [ ] Pattern similarity clustering (catch near-duplicates)
• [ ] Team export/import (skill-fog export --team)
• [ ] VS Code extension
• [ ] Configurable thresholds per project

Contributing

PRs and issues are welcome. This is a bash-first project — keep it simple.

1. Fork: github.com/pabang0620/skill-fog
2. Branch: git checkout -b feature/my-feature
3. Commit: git commit -m 'feat: add my feature'
4. Push: git push origin feature/my-feature
5. Open a Pull Request

If you have a pattern that skill-fog should detect better, open an issue with a real example. The normalization logic lives in hooks/stop.sh and is easy to extend.

License

MIT © pabang0620