Coordinate parallel independent tasks with evidence sharing and pass/fail criteria.
npx skills add https://github.com/Orinks/AccessiWeather --skill ultrawork-orinks
Please help me install this Agent Skill. Skill: ultrawork Source: https://github.com/Orinks/AccessiWeather/tree/main/.codex/skills/ultrawork Command: npx skills add https://github.com/Orinks/AccessiWeather --skill ultrawork-orinks
Ultrawork addresses the inefficiency of sequential task execution by providing a parallel execution engine that coordinates multiple tasks, maintains context discipline, and offers smart delegation guidance without enforcing persistence or long-running guarantees.
Start Ultrawork on a batch of independent tasks and observe parallel progress.
High-intent search queries and answers about installing and using this skill.
You run multiple independent tasks concurrently using a parallel execution engine that coordinates workflows while preserving context discipline and traceability. It maintains lightweight evidence lanes for progress tracking.
A two-track execution model with direct-tool and evidence lanes is a structured parallel processing approach where one track executes tools directly and the other generates verifiable evidence. This ensures results are verifiable without full persistence guarantees.
You verify parallel task results without full persistence guarantees by using lightweight evidence lanes that attach evidence as tasks progress. This satisfies functional requirements for explicit acceptance criteria and clear pass/fail verification.
You need explicit acceptance criteria for concurrent task management when running workflows where many tasks execute in parallel and results must be verifiable. This ensures clear pass/fail criteria are met across independent parallel lanes.
You can integrate parallel execution with persistence layers like Ralph or autopilot when a verified completion lifecycle is required. This provides lightweight integration for optional persistence alongside the concurrent execution engine.
The limitations of parallel task execution without long-running guarantees include a lack of full persistence and no enforced long-running stability. It is designed for lightweight evidence sharing and smart delegation rather than durable pipeline lifecycles.
wmasfoeOrchestrate parallel execution of independent tasks with lightweight evidence and acceptance criteria.
RJJJJLoop tasks until completion with architect verification and session persistence.
BetaSExecute persistent task loops with parallel agent calls and architect verification.
frankxaiCoordinate parallel task execution across agents and tools with thread-safe dispatch and failure recovery.
