sisyphus

Coordinate multi-step software workflows with delegation, MECE planning, and evidence-audit gating.

25|1|Updated Jan 14, 2026
One-click install
npx skills add https://github.com/toongri/oh-my-toong-playground --skill sisyphus-toongri
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: sisyphus
Source: https://github.com/toongri/oh-my-toong-playground/tree/main/skills/sisyphus
Command: npx skills add https://github.com/toongri/oh-my-toong-playground --skill sisyphus-toongri

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Orchestrates complex multi-step tasks that would otherwise tempt a single agent to do everything, enforcing delegation, parallelization, and QA ownership so the workflow never bypasses specialists or unverified changes.

Core Features & Use Cases

  • Delegation enforcement encodes RULES 1–4 and the agent routing table so every change touches a specialist, no code edits happen directly, and argus verification is always invoked.
  • Complexity gating relies on the MECE task planner, work-unit slugs, and parallelization analysis to split goals into atomic delegations, track blockedBy links, and surface dependency-aware work batches.
  • Verification discipline executes the full Task Execution Loop, Evidence Audit Gate, and Verdict Response Protocol so each stratum from junior to mnemosyne runs sequentially with traceable evidence.

Quick Start

Ask sisyphus to orchestrate your next multi-step change, create MECE tasks, assign specialists, and reserve argus for verification.

Frequently Asked Questions about sisyphus

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
What is multi-agent task orchestration for complex software engineering workflows?

Multi-agent task orchestration coordinates complex multi-step software engineering workflows by splitting goals into atomic delegations, enforcing strict agent routing, and using evidence-backed QA to verify every code change.

How do I delegate tasks across multiple agents for software architecture and debugging?

Delegate software architecture and debugging tasks by applying MECE task planning, generating work-unit slugs, analyzing parallelization, and routing each atomic work unit to a specialist agent based on a strict routing table.

How does evidence-backed QA verification work for multi-agent code changes?

Evidence-backed QA verification executes a Task Execution Loop and Evidence Audit Gate, requiring each delegated code change to pass specialist verification sequentially with traceable evidence before final approval.

When should I use agent routing and task orchestration instead of a solo agent?

Use agent routing and task orchestration when complex multi-step software engineering changes risk overwhelming a solo agent, requiring read/write division, parallelization analysis, and strict verification discipline.

Can I parallelize multi-step workflows with dependency tracking using task orchestration?

Yes, task orchestration parallelizes multi-step workflows by applying a MECE task planner to analyze dependencies, track blockedBy links, and surface dependency-aware work batches for parallel agent execution.

What are the limitations of using strict delegation rules for agent routing?

Strict delegation rules prevent direct code edits and require specialist invocation for every change, which may introduce overhead for simple, single-step modifications that do not require complex multi-step orchestration.