Claude Skills

A small, curated collection of Agent Skills for Claude Code (and any skill-compatible agent). Each skill is a self-contained folder — drop it into ~/.claude/skills/ and Claude can use it.

Everything here is bring-your-own-credentials: no API keys, tokens, or personal data are stored in the skills. Each skill's SETUP.md walks you through connecting your own accounts.

Skills

🔎 local-lead-prospector — Local business lead prospector

Finds and scores local businesses that need a website (or a better one), for a web-design / digital-services business. It ranks leads by need (no site / Facebook-only / outdated) × ability to pay (Google review volume, sector, scale), tags each with a €-budget tier and a ready-to-say call hook, and writes them to a Notion table as a call-then-email work list. Ireland-first, token-efficient, and portable to other countries.

• Discovery via Golden Pages (free), optional Google Places enrichment for real review data.
• Scoring, €-budget, hooks and the Notion write live in a reusable prospect.py; enrich.py pulls live Google reviews and re-scores.
• See its SETUP.md.

✉️ cold-email-drafter — Cold-email drafter

Turns a scored lead from the prospector into a short, human cold email with zero LLM tokens — one editable template is filled from the lead's data by a script. Can create the email as a Gmail draft (never sends) or a one-click Gmail compose link you press from Notion. Pairs with the prospector.

Install a skill

# copy the skill folder into your Claude skills directory
cp -R skills/local-lead-prospector ~/.claude/skills/

Then open its SETUP.md and connect your own accounts (Notion; optionally Google Places and Gmail). Invoke it in Claude Code with /local-lead-prospector.

Repository layout

skills/
  <skill-name>/
    SKILL.md        # the skill definition (required)
    *.py            # any helper scripts the skill uses
    SETUP.md        # optional: one-time setup / credentials

Add your own skill

1. Create a folder under skills/ named for your skill.
2. Add a SKILL.md with YAML frontmatter (name, description, allowed-tools) and instructions.
3. Keep all secrets out of the folder — read them from environment variables or a file in ~/.claude/ (see how the existing skills do it), so the skill stays safe to share.
4. Add a SETUP.md if it needs credentials.

Safety

• No API keys, tokens, database IDs, or personal data are committed. Credentials load from ~/.claude/ config files or environment variables at runtime.
• .gitignore blocks common secret files as a backstop.

License

MIT — see LICENSE.