Fan out independent tasks to Codex and Sonnet workers in isolated environments.
npx skills add https://github.com/yzhao062/anywhere-agents --skill prun
Please help me install this Agent Skill. Skill: prun Source: https://github.com/yzhao062/anywhere-agents/tree/main/packages/pypi/anywhere_agents/composer/skills/prun Command: npx skills add https://github.com/yzhao062/anywhere-agents --skill prun
This Skill solves the bottleneck of sequential task execution by enabling massive parallel delegation, allowing you to fan out complex projects into independent units that run concurrently across multiple worker types.
Use the prun skill to decompose the current task into independent units and dispatch them to parallel workers for execution.
High-intent search queries and answers about installing and using this skill.
To run parallel code generation and bypass quota constraints, you can fan out independent task units to separate Codex workers. This orchestration ensures heavy workloads execute concurrently without stalling your primary session.
Parallel task delegation decomposes large projects into independent units dispatched concurrently across multiple worker types. It enables high-throughput code generation and research by orchestrating execution across specialized Codex and Sonnet workers.
To safely integrate parallel code generation outputs, workers operate in isolated, throwaway clones. This safety model ensures your primary repository remains untouched until you manually review and reconcile the multi-worker results.
Yes, you can use Codex and Sonnet workers together for concurrent task orchestration. The system automatically dispatches independent units to either worker type based on specific capability requirements for code generation or analysis.
If a parallel worker stalls during task execution, automated monitoring and stall detection mechanisms identify the issue. The system actively manages complex workflows to ensure reliable reconciliation of multi-worker outputs.
You should not use parallel delegation for sequential or tightly coupled software engineering tasks. This approach requires independent units of work to fan out effectively, making interdependent workflows unsuitable for concurrent worker dispatch.
NQC233Dispatch independent agents to execute parallel tasks in isolated contexts.
aryan-binazirDelegate implementation, analysis, or review tasks to worker models and output JSON results.
harshaldhadukBenchmark agentic-flow worker operations against configurable latency thresholds.
phuonghxOrchestrate multi-agent task execution with decomposition, parallel reads, and sequential writes.
