Loop a task with persistent retries until architect verification completes.
npx skills add https://github.com/shaun0927/codex-superskills --skill until-done
Please help me install this Agent Skill. Skill: until-done Source: https://github.com/shaun0927/codex-superskills/tree/main/skills/ralph Command: npx skills add https://github.com/shaun0927/codex-superskills --skill until-done
Ralph creates a self-persisting loop that keeps working on a task until it is fully complete and architect-verified.
Initiate a Ralph loop on the current task, enforcing architect verification before completion.
High-intent search queries and answers about installing and using this skill.
To keep a task looping until true completion, you can use an iterative persistence mechanism that enforces architect verification before closure. This approach maintains state across retries and delegates parallel subtasks automatically.
Automating persistent iteration across task retries requires a self-persisting loop that continuously works on the objective. It ensures tasks spanning multiple iterations are fully executed and verified before finalization.
Yes, you can delegate parallel subtasks to specialists with automatic retry while requiring explicit verification before the main loop exits. This ensures all delegated components are fully complete before closure.
To chain task iterations and cleanly exit an automation loop, use a workflow that verifies progress at each step. You can initiate the loop on the current task and execute a clean cancel once completion is verified.
A multi-step task closes prematurely if it lacks mandatory architect verification before finalization. Implementing a persistent loop with explicit completion checks ensures parallel delegation is fully verified before exiting.
ozbyOrchestrate persistent task execution with retry and architect verification.
taku-oVerify task completion claims against fresh validation evidence and exit codes.
dev-khoiRun independent LLM subtasks concurrently and merge results with deterministic logic.
wunkiExecute PLAN.md tasks sequentially with human approval and learning persistence.
