runesleo

Project knowledge accumulation system - learn from practice, avoid repeating mistakes

File-based planning for complex tasks. Creates task_plan.md, findings.md, and progress.md. Use for multi-step tasks requiring >5 tool calls.

Session wrap-up - update handoff + commit + auto-record experience

Systematic debugging - four-phase process, find root cause before fixing

Verification before completion - must run verification commands before claiming done. Evidence before claims.

Precise PnL calculator for any Polymarket address — reconstructs profit/loss from Data API cashflow (BUY/SELL/REDEEM/MERGE/SPLIT/REBATE) plus unrealized position value. Matches official /profit within ~0.2% MAPE on leaderboard validation. Bring-your-own Python (httpx).

Brier Score calculator for Polymarket addresses — measures prediction accuracy independent of PnL. Separates skilled predictors from market makers and arbitrageurs.

Polymarket address profiler — input any 0x address, get a complete trading profile with PnL, win rate, positions, category breakdown, and top trades. All data from public APIs, no local database needed.

扫描 ~/.claude/skills/ 输出热度/死 skill/冲突/过期 pattern 报告。周日运营日跑。触发：/skill-audit、"技能树体检"、"skill 健康度"、"哪些 skill 没用"。

Four-phase debugging framework that ensures root cause investigation before attempting fixes. Never jump to solutions.

4-stage content pipeline orchestrator: Research -> Ideate -> Write -> Queue. Give it a topic, it researches existing discussions, generates hook angles, writes a draft, and queues it for review. Inspired by @shannholmberg's 4-Agent content system (Research -> Ideate -> Write -> Orchestrate). Designed for creators who build in public and want systematic content production.