İçeriğe geç
TakoAPI
Becerilere dön

intelligent-delegation

hogpile tarafından

Search & Research
436İndirme sayıları clawskills.sh tarafından bildirilir
0Görüntülenmeler
0 beğeniBu beceriyi beğenmek için giriş yapın
—Henüz değerlendirilmedi
Puanınız

Özet

A 5-phase framework for reliable AI-to-AI task delegation, inspired by Google DeepMind's "Intelligent AI.

Açıklama

Bu Beceri Ne Yapar

A 5-phase framework for AI-to-AI task delegation based on Google DeepMind's Intelligent AI Delegation paper (arXiv 2602.11865). Covers task tracking, sub-agent performance logging, automated output verification, fallback chains, and systematic agent selection via multi-axis scoring.

Replaces ad-hoc delegation with a structured, self-correcting system that logs outcomes, enforces contracts, and routes failures automatically rather than silently dropping them.

Ne Zaman Kullanılır

  • Tracking background sub-agent tasks so none go unmonitored
  • Choosing which sub-agent to assign a task based on past performance
  • Automatically verifying sub-agent output before accepting results
  • Recovering from delegation failures without manual intervention
  • Scoring task complexity and reversibility to decide autonomy level

Örnek İş Akışı

GİRİŞ

User asks: delegate a data extraction task to a sub-agent and ensure results are valid

AJAN
  1. 1

    Score the task on 7 axes using score_task.py to determine autonomy level and agent choice

  2. 2

    Write a task contract specifying expected output format, success criteria, and fallback steps

  3. 3

    Log the task to TASKS.md and schedule a cron job to check completion

  4. 4

    On completion, run verify_task.py to confirm file exists and passes structural checks

  5. 5

    If verification fails, apply the fallback chain: retry with adjusted prompt, then try a different agent, then escalate

ÇIKIŞ

Verified output file at expected path, task marked COMPLETED in TASKS.md, outcome logged to agent-performance.md

Kurulum

clawhub install hogpile/intelligent-delegation

CLI komutunu çalıştırın ya da istemi Claude Code, Cursor, Codex veya kullandığınız kodlama ajanına yapıştırın.

Bağlantılar