Iterate on software engineering tasks with TODO tracking and verification.
npx skills add https://github.com/realsigridjin/oh-my-codex --skill ralph-realsigridjin
Please help me install this Agent Skill. Skill: ralph Source: https://github.com/realsigridjin/oh-my-codex/tree/main/skills/ralph Command: npx skills add https://github.com/realsigridjin/oh-my-codex --skill ralph-realsigridjin
Ralph prevents complex engineering tasks from being abandoned halfway or declared complete without fresh verification, keeping work persistent across iterations until an architect confirms the result.
Activate the ralph skill and provide the engineering task you want completed until it passes verification and receives architect approval.
High-intent search queries and answers about installing and using this skill.
To ensure software engineering tasks are fully completed, use persistent iteration with mandatory architect verification. This approach tracks TODO items across retries and requires fresh test and build evidence before declaring completion.
The best way to manage multi-step feature implementation is parallel delegation, routing independent tasks to specialized agents using appropriate capability tiers. This routes independent tasks to specialized agents while tracking prior progress.
Persistent agent orchestration handles incomplete work by continuing iterations across retries while tracking TODO items and prior progress. It revisits rejected work and handles recurring blockers until an architect confirms the result.
Yes, you can use agent-tier routing for multi-file feature implementation requiring architectural review. It routes independent tasks to specialized agents while requiring architect approval and clean state cleanup for reliable continuation.
Automated engineering tasks stop before passing verification when they lack persistent iteration and mandatory architect approval. Requiring fresh diagnostic evidence and TODO tracking ensures work continues until fully complete and verified.
Yes, Ralph requires fresh test, build, and diagnostic evidence before task completion. It mandates architect-review evidence and clean state cleanup, ensuring complex implementations pass verification rather than being abandoned halfway.
Ethical-AI-SyndicateGuide OKR creation with structured objectives, key results, and YAML output.
mazenyassergithubRoute iterative development tasks to specialized agents with architect verification.
oneforceCoordinate parallel task execution across multiple agents with tier-aware routing.
sigridjinethOrchestrate iterative task completion with parallel agent delegation and architect verification.
