惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

腾讯CDC
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
博客园 - 叶小钗
人人都是产品经理
人人都是产品经理
博客园 - 聂微东
The Cloudflare Blog
爱范儿
爱范儿
阮一峰的网络日志
阮一峰的网络日志
WordPress大学
WordPress大学
小众软件
小众软件
博客园 - 三生石上(FineUI控件)
Last Week in AI
Last Week in AI
Jina AI
Jina AI
V
V2EX
罗磊的独立博客
V
Visual Studio Blog
A
About on SuperTechFans
IT之家
IT之家
P
Proofpoint News Feed
B
Blog
博客园 - Franky
Blog — PlanetScale
Blog — PlanetScale
Google DeepMind News
Google DeepMind News
Y
Y Combinator Blog

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant
Bounty Watcher: An Autonomous AI Agent That Finds Paid Gi...
Alexandre La · 2026-05-16 · via DEV Community

Two shell scripts. A cron scheduler. Hermes Agent as the delivery layer. Zero cloud costs. Freelance opportunities delivered to your Telegram every few hours.

What I Built

Bounty Watcher is an autonomous agent pipeline running on Hermes Agent that scans freelance bounty platforms 24/7, filters out noise, deduplicates, and delivers only relevant opportunities — straight to Telegram.

It covers two platforms:

  • Superteam Earn — crypto/solana bounties with cash rewards
  • GitHub Issues — open source bounties across the ecosystem

The agent has been running in production for weeks on an Android phone via Termux. It costs nothing to operate beyond the LLM tokens used for occasional reasoning tasks — which, in this architecture, is close to zero.

Category Submission

Build With Hermes Agent — Hermes Agent is the orchestrator: it schedules, monitors, delivers, and provides the agentic intelligence for filtering and prioritization.

How Hermes Agent Powers This

Architecture

+---------------------------------------------+
|               Hermes Agent                   |
|                                              |
|  +----------+  +----------+  +-----------+  |
|  |  Cronjob |  |  Cronjob |  |  Memory   |  |
|  |  (6h)    |  |  (3h)    |  | (dedup)   |  |
|  +----+-----+  +----+-----+  +-----------+  |
|       |             |                        |
|  +----v-----+  +----v-----+                 |
|  |Superteam |  |  GitHub   |                 |
|  | Script   |  |  Script   |                 |
|  +----+-----+  +----+-----+                 |
|       |             |                        |
|       +------+------+                        |
|              v                               |
|       +-------------+                        |
|       |  Delivery   |                        |
|       |  (Telegram) |                        |
|       +-------------+                        |
+---------------------------------------------+

Enter fullscreen mode Exit fullscreen mode

Hermes Agent plays three roles:

  1. Scheduler — The cronjob tool runs shell scripts on a fixed cadence (every 6h, every 3h). The scheduler handles retries, timezone management, and deterministic delivery routing.

  2. Filter — Scripts use relevance scoring with curated keyword lists (Solana, DeFi, Rust, Python, bounties, etc.) to surface only matching opportunities. The scripts maintain a local cache to track seen issues and highlight new ones.

  3. Messenger — Output is delivered directly to Telegram. When a script finds nothing new, it produces empty output — and Hermes Agent's cron system sends nothing. Silent by design. No spam.

Why No-Agent Crons?

Both bounty watchers use no_agent: true. This is a deliberate design choice:

  • Zero LLM cost — scripts run purely in shell/Python, no tokens consumed for data collection
  • Deterministic — same input always produces the same filtered output
  • Fast — execution completes in under 2 seconds per scan
  • Silent when empty — the watchdog pattern: script outputs text only when there's something to report

The Superteam Scanner

This script queries the Superteam Earn API, filters bounties by agentAccess status, and prioritizes by deadline proximity:

# Core filtering logic
RELEVANCE = [
    'solana', 'agent', 'defi', 'crypto', 'trading', 'blockchain',
    'smart contract', 'usdc', 'usdt', 'API', 'web3', 'rust', 'python',
    'typescript', 'nft', 'token', 'dex', 'amm', 'oracle', 'bridge',
    'validator', 'staking', 'airdrop', 'spl', 'jupiter', 'raydium',
    'sdk', 'CLI', 'cli tool', 'automation', 'bot', 'scraper',
    '$', 'reward', 'prize', 'earn', 'grants',
]

SKIP = ['translation', 'translate', 'typo', 'watchlist', 'parity']

# Relevance scoring
all_text = title_lower + ' ' + ' '.join(labels)
score = sum(1 for kw in RELEVANCE if kw.lower() in all_text)

if score == 0:
    continue  # Skip irrelevant

Enter fullscreen mode Exit fullscreen mode

Key features:

  • Deduplication via persistent cache
  • AGENT_ALLOWED filtering — highlights bounties that explicitly permit AI-assisted submissions
  • Deadline awareness — sorts by hours remaining, skips expired
  • Human-only overflow — shows high-value human-only bounties ($1000+) as context

The GitHub Scanner

Uses GitHub CLI to search for bounty-labeled issues, then applies the same relevance engine.

Real Results

After weeks of running, the watchers have surfaced:

  • Superteam: 15-25 active bounties per scan, 3-5 with AGENT_ALLOWED status
  • GitHub: 5-10 relevant issues per scan, filtered from 30+ raw results
  • Signal-to-noise: ~30% of fetched results pass the relevance filter
  • Novelty detection: 2-4 new bounties per day

How to Deploy It Yourself

Prerequisites

  • Hermes Agent installed (pip install hermes-agent)
  • GitHub CLI (gh) authenticated
  • A Telegram bot connected to Hermes Agent

Step 1: Clone the repo

git clone https://github.com/AtlasNexusOps/hermes-bounty-watcher.git
cp hermes-bounty-watcher/scripts/* ~/.hermes/scripts/

Enter fullscreen mode Exit fullscreen mode

Step 2: Schedule the crons

hermes cron add "every 360m" \
  --name "Superteam Bounties" \
  --script ~/.hermes/scripts/superteam_bounties.sh \
  --no-agent \
  --deliver origin

hermes cron add "every 180m" \
  --name "GitHub Bounty Scan" \
  --script ~/.hermes/scripts/github_bounties.sh \
  --no-agent \
  --deliver origin

Enter fullscreen mode Exit fullscreen mode

Step 3: Let it run

That's it. You'll start receiving formatted bounty digests in Telegram every few hours.

What Makes This a Strong Hermes Agent Build

  • Hermes Agent is the backbone, not just a wrapper
  • Production-proven on modest hardware (Android/Termux)
  • Practical value — finds real paid opportunities
  • Composable — swap API endpoints and keywords for any platform
  • Cost-efficient — no-agent crons = zero LLM tokens for data collection

Future Directions

  • Multi-platform expansion (Gitcoin, OnlyDust, Bountycaster)
  • Agent-powered daily briefings
  • Automated proposal drafting

Built with Hermes Agent, running on an Android phone, delivering value every day.