Specification-driven development toolkit for DeepSeek Harness.
Recause is an experimental programming model and runtime for writing interactive software as ordinary code. Flows stop when information is missing, represent progress in explicit serializable state, and run again when new state arrives—without callbacks, explicit resume logic, or hand-authored state machines.
Protocol for durable, resumable, evidence-backed work across long-running agents, humans, and tools.
Durable manual-core workflow skill for recursive planning, review/fix loops, and resumable AI-agent execution.
Clean-room Python reference for approval-gated, resumable bulk-file operations with synthetic fixtures and duplicate-aware tests.
Run record, resume decision, and reuse-history primitives for FlowBrigade-style workflows.
Agent skill for building local, resumable AI pipelines.
Local, resumable AI pipelines. A little TypeScript and a SQLite database.
Build complete intelligent systems in Codex: continuous requirements grill, adaptive exploration, software and intelligence architecture, implementation and verification. 在 Codex 中持续探索、编排与实现完整智能系统；一个总控与五个模块化 Skill。
Agent-native workflow for translating full books with durable state, source-grounded review, resumable progress, and multi-book workspaces. Works with ChatGPT, Codex, and other AI agents.
Codex skill for resumable migrations, refactors, debugging, and evidence-backed completion. Planned support for Paperclip.
To associate your repository with the resumable-workflows topic, visit your repo's landing page and select "manage topics."