
Yora 又部
170 posts






built an autonomous Base trading agent running 5 parallel loops continuously. 1. Scanner Loop - every 5 minutes the core engine. * pulls 20 trending Base tokens from GeckoTerminal * filters out: * liquidity under $50k * tokens dumping >15% in 1h * scores every token from 0–100 using: * dip depth * bounce confirmation * sentiment shift * buy pressure * volume quality * higher timeframe trend if score ≥ 60: → runs GoPlus rug check → auto buys through Uniswap v3 2. Monitor Loop - every 10 seconds manages every open position automatically: * stop loss: -6% * take profit: +25% * trailing stop: * activates at +4% * trails 3% below peak * max hold: 45 minutes fully deterministic. no emotions involved. 3. Heartbeat Loop - every 5 minutes creates system snapshots: * ETH balance * open positions * unrealized PnL * active risk exposure the LLM only wakes up if: * wallet balance gets too low * a position behaves abnormally * execution fails keeps inference costs low. 4. Agent Loop - every 90 minutes the LLM reviews market conditions and dynamically changes trading behavior: * active → auto buys best candidates * selective → entries require LLM approval * watchOnly → scans market without trading 5. Reflect Loop — every 4 hours the self-improvement cycle. the LLM reviews: * closed trades * win/loss patterns * bad entries * volatility conditions then automatically adjusts: * minimum score thresholds * liquidity requirements * risk configs * position sizing the agent slowly adapts to market conditions over time. full trade flow: GeckoTerminal data → scoring engine → GoPlus rug check → Uniswap v3 execution → live monitoring every 10s → auto sell via SL/TP/trailing → reflection + strategy adaptation most important part: the trading engine itself is deterministic. LLM is NOT making every trade decision. it’s only used for: * reflection * market adaptation * exception handling * operator chat execution stays: fast, cheap, deterministic, and reliable. open source. run your own: github.com/noelclaw/noelc…




we’ve submited our app to the @gitlawb playground. kindly approve bruh @kevincodex


Whoaaah OpenGateway top 4 in global ranking of OpenRouter! Did we even surpass Claude Code? 😳😳😳



We are buiidng automated security scanning for gitlawb & gitHub. gitscan scans @gitlawb and @github repos for secrets, vulnerabilities, and CVEs — then explains everything in plain English. TruffleHog detects leaked API keys, tokens, and private keys in your commit history. Semgrep analyzes code patterns to surface injection flaws and security anti-patterns. CA : 0x46BC5B1b003e9659d5638715e3302e15C372d59d Website : gitlawbscan.com Github : github.com/gitlawbscanner… tg bot : t.me/gitscanbot






