Install the codex-dynamic-workflows skill so an AI agent can plan and run a multi-step task with approval checkpoints
Break a large task into subagents or simulated work packets that a supervised agent workflow tracks
Reuse a completed workflow's artifacts as a starting point for a similar future task
| dannymac180/skills | zhengdian1/interleavethinker | fangcun-ai/skillward | |
|---|---|---|---|
| Stars | 124 | 124 | 123 |
| Language | Python | Python | Python |
| Setup difficulty | easy | hard | moderate |
| Complexity | 2/5 | 5/5 | 4/5 |
| Audience | developer | researcher | ops devops |
Figures from each repo's GitHub metadata at analysis time.
This repository is a small public collection of skills for AI coding agents, published by a developer who goes by DannyMac180. A skill in this context is a packaged set of instructions that an AI agent can load and follow to handle a particular kind of task, rather than figuring out the approach from scratch each time. At the moment the collection holds one skill, called codex-dynamic-workflows. According to its description, it lets an AI agent plan and run supervised, multi-step workflows. That includes a goal mode where the agent works toward an outcome, the option to use subagents or simulated work packets to break a task into pieces, approval gates so a human can check in before the agent continues, and steps for integrating and verifying the work once it is done. It also produces reusable workflow artifacts, meaning the plan or output from one run can be kept and used again later. Installing a skill is simple. If the AI agent a person is using already supports skills, they can just point it at the skill's GitHub URL and ask it to install that skill directly. Alternatively, someone can clone the whole repository, then copy the specific skill folder into their agent's own skills directory. The README gives an example for Codex, copying the codex-dynamic-workflows folder into the Codex home directory's skills folder. Once installed, the skill is invoked by name in a new agent session, followed by a description of the task to hand off to it. The README itself is brief and does not go into detail about how the workflow logic is implemented internally. The project is released under the MIT license.
A small collection of AI agent skills, currently containing one skill that lets an agent plan and run supervised, multi-step workflows with approval checkpoints.
Mainly Python. The stack also includes Python.
Use, modify, and distribute freely, including for commercial purposes, as long as you keep the original copyright notice.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.