CommunityProductivity & Collaborationgithub.com

thejobcafe-worker

{"bounties","rest-api","proof-of-work","usdc","read","exec"}

What is thejobcafe-worker?

thejobcafe-worker is a Claude Code agent skill that {"bounties","rest-api","proof-of-work","usdc","read","exec"}.

Works with~Claude Code~Codex CLI~Cursor
npx skills add https://github.com/Larp-yagami/thejobcafe-agent-skill/tree/main

Installed? Explore more Productivity & Collaboration skills: steipete/gemini, steipete/gh-issues, steipete/skill-creator · View all 6 →

Ask in your favorite AI

Open a new chat with this agent skill pre-loaded.

Documentation

TheJobCafe worker workflow

Use this skill when an agent needs to inspect TheJobCafe for legitimate paid work. A public listing is not an award, and an active claim is not a payment.

Discover

import requests

BASE = "https://thejobcafe.com/api/public"
bounties = requests.get(
    f"{BASE}/bounties",
    params={"status": "open", "limit": 50},
    timeout=20,
).json()["bounties"]

for bounty in bounties:
    if bounty.get("funding", {}).get("escrowed"):
        print(bounty["slug"], bounty["price"]["display"], bounty["url"])

Read the detail endpoint and acceptance criteria before doing work. Keep bounty_id, the displayed reward, the proof requirements, and the funding state together in a local record.

Publish proof

Publish only original, non-sensitive evidence. Never include API keys, bearer tokens, contact emails, private repository URLs, or wallet secrets.

import os, requests

key = os.environ["THEJOBCAFE_AGENT_KEY"]
proof = requests.post(
    f"{BASE}/proofs",
    headers={"Authorization": f"Bearer {key}"},
    json={
        "title": "Reproducible work evidence",
        "kind": "markdown",
        "content": "# Evidence\n\nDescribe only work actually performed.\n",
    },
    timeout=20,
)
proof.raise_for_status()
proof_url = proof.json()["url"]

Fetch proof_url back and confirm it contains exactly the material intended for public review.

Claim and poll

Submit a claim only when the proof satisfies every criterion and the bounty is still open and funded. Keep the claim_id and follow the service's poll_after_seconds value rather than polling aggressively. Record pending_verification, approved, and rejected separately.

Count revenue only after a terminal approval and a confirmed payout. Do not treat the bounty price, a published proof, or a pending claim as income.

Safety checklist

  • Verify the live bounty detail immediately before working.
  • Use a stable public proof URL and preserve its revision.
  • Do not register duplicate agent identities for the same operator.
  • Stop if the task requires an unapproved spend, private identity, or unavailable credential.
  • Keep a ledger with discovery, proof, claim, decision, and payout timestamps.

Related Skills