Skip to content

Jobhunt Docs

Living documentation for the jobhunt project - scrapers, scorer, references, strategy.

What's in here

Architecture

  • Architecture - end-to-end flow with Mermaid diagrams, plus the gaps between "Claude in chat" and "the agent".
  • Data Sources Catalogue - every source, what's in the tile vs detail-fetch, auth, anti-bot, cadences.

Criteria (the rules the scorer reads at runtime)

  • Upwork criteria - Hustle landing-page lane + occasional Jonny AI/MVP gigs.
  • Non-Upwork criteria - Jonny's PM career (perm + contract). Founder-gap, fintech subdomain distinctions, requirement-pass-rate rule, location default, EOR.
  • Portfolio inventory - Hustle/Jonny's portfolio + positive/reject keyword lists.

References (system prompts for Lovable + the client-research skill)

Upwork strategy

  • README - how the strategy folder is organised.
  • Playbook - bidding rules (headline, timing, structure, sign-off, rejects).
  • Packages - $499 / $1,499 / $2,999 tier specs + add-ons.
  • Niche portfolio - 7 spec pages to build (verticals × tiers).
  • Todo - phase-ordered backlog.
  • Myths debunked - claims discarded after data.
  • Tests in flight - A/B tests running on real bids.

Where edits go

These docs are rendered from the source .md files in the repo:

Tab Source path
Architecture, Data Sources docs/architecture.md, references/
Criteria config/criteria_*.md, config/portfolio.md
References references/*.md
Upwork Strategy strategy/upwork/*.md

Edit the source .md files in the repo, commit, and the docs site auto-rebuilds (or run mkdocs serve locally to preview).