Packmind Open Source
Roll out coding standards safely across repos and agents
145 followers
Roll out coding standards safely across repos and agents
145 followers
AI assistants code fast — but without strong and shared context, they amplify inconsistency. Packmind OSS helps teams create, scale, and govern the engineering playbook that keeps AI coding safe, consistent, and aligned with your standards.
Packmind Tech Coach
👋 Hey Product Hunt,
I’m Cédric, CTO and co-founder of Packmind. We’ve just built an open-source framework for Context Engineering — versioning, distributing, and enforcing organizational standards across repos and coding agents.
### Why we built it
Over the past year, we’ve been scaling AI-assisted development across our teams. Today, about 65% of our commits come from coding agents (Copilot, Cursor, Claude Code, and Kiro).
The productivity gain is real — but so is the drift it creates.
Each assistant ends up coding from a different context snapshot of our architecture, naming conventions, and standards. Some pull from outdated instruction files, others from old wikis. The result: AI-generated code that’s locally correct but globally inconsistent.
We built Packmind OSS to fix that.
### What Packmind OSS does
Create, scale, and govern your engineering playbook for AI coding assistants (Copilot, Cursor, Claude Code, Codex, Kiro…).
✅ Create — turn scattered rules from wikis, ADRs, and code reviews into a living playbook.
✅ Scale — auto-sync the same context across all repos & agents.
✅ Govern — check adherence, visualize drift, and repair it automatically.
### 🔥 We’d love your thoughts
- What’s been most frustrating about managing the context for your AI coding agents?
- How do you keep standards and prompts consistent across repos or assistants?
- What features or integrations would you love to see next?
We’re early, learning fast, and curious about how other teams are scaling AI-assisted development safely.
👉 OSS Repo: https://github.com/PackmindHub/packmind
Cal ID
Congrats on the launch! Having consistent coding standards across AI agents is such an important challenge to solve.
How does Packmind handle conflicts when different teams want different coding standards for the same repository?
Packmind Open Source
@sanskarix Hello Sanskar, that’s a great question! Right now, teams can define both their own specific standards and share common ones across the organization. For example, you might have shared standards for security or Java, while individual teams can define additional rules based on their local context.
We’re also working on a feature that will alert teams when standards conflict or overlap, so they can align and rationalize them at the organization level. 🙂
Cal ID
@arthur_magne The alerting team feature sounds interesting. Thanks!
This is super relevant — “locally correct but globally inconsistent” perfectly sums up the pain point of scaling AI-assisted dev. 👏
Curious: how does Packmind handle context drift detection between repos? Is it diff-based or embedding-based?
R2Devops
Congrats on the launch! Drift in the code of all these agents is a big topic to solve.
I'm glad to see you open-source Packmind and curious to have your feedback on this big step ;)
Packmind Tech Coach
@thomas_boni Thanks for your message Thomas :-)
Love this! Keeping AI coding consistent across teams is tough. How’s the community using Packmind to align agents with their coding standards?
Packmind Tech Coach
@mohammed_maaz3 Thank you for your feedback! Since the launch, we've gone from 0 to 80+ stars in less than a week. The initial feedback has been great. You can check out the repo, see the PRs, and follow future conversations with the community here: https://github.com/PackmindHub/packmind