How to organize work in GitHub
A developer uses GitHub repositories, scoped folder instructions, Copilot, and agents to coordinate coding and operational work, including AI summaries of goals and progress.
news.ycombinator.comSuggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Using [repository] and the instructions in [folder scope], turn [idea or operational task] into a specification and actionable work. Log progress against [goals and KPIs], then summarize achievements and outstanding work for the team.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source poster / publisher
- dipeshsukhani
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- The author reports shipping small open-source tools, shorter delivery loops, and largely replacing stand-ups with AI-mediated updates; the 10–20x estimate is subjective.
- Human involvement
- The post does not describe specific approval or review checkpoints for agent output.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- The author organized coding and operational work into GitHub repositories and folders.
- They added global and scoped context and instructions for agents.
- They routed idea, specification, implementation, and iteration work through AI.
The author reorganized their work around GitHub repositories containing both software projects and operational activities. They report using Copilot and other agents inside those repositories rather than relying mainly on standalone chatbot conversations. Their repositories cover hiring, application review, financial reviews, and other workflows alongside code.
The structure uses master repositories with folders and subfolders. Each area contains its own context and instructions, giving agents global rules plus more specific guidance for the scope where they operate. The author connects that organization to shorter loops from idea to specification, implementation, and iteration. They also report shipping a couple of small open-source tools they believe they would otherwise not have finished.
For team coordination, KPIs and goals live in repositories, while progress and achievements are logged and summarized through AI. The team has largely replaced traditional stand-ups with these updates. The author describes tasks shrinking from weeks to days or hours, but explicitly presents the broader 10–20x impression as subjective and difficult to measure precisely.
Who it's for: Teams already using GitHub for shared context may find scoped repository instructions useful for organizing agent-assisted work.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: Hacker News · dipeshsukhani
Frequently asked questions
How are instructions scoped to different kinds of work?
The author uses master repositories with folders and subfolders containing specific context and instructions, alongside global rules.
Is the claimed 10–20x productivity increase measured?
The author calls it a subjective feeling and says it is hard to measure precisely, while reporting that some tasks now take days or hours instead of weeks.
More details
Remaining steps
- The team logged goals and progress in repositories and used AI to summarize updates.
About this account
Copilot and repository agents support work across scoped areas, including implementation and progress summaries.
Platform: Hacker News. Attribution: https://news.ycombinator.com/item?id=46409668. Prompt: Suggested (written by Agents In Use, not the author's). Sourced by the Agents In Use pipeline.
Credit: dipeshsukhani on Hacker News
Source published
Added to Agents In Use
Source checked
Tools and topics
GitHub · Copilot
GitHub · Team updates · Scoped instructions
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.