AI PACKS for pro@ AI Pack Namespaces

This is the repository for Production Coding Packs, developed and maintained by Jeremy Chone and the AIPACK & BriteSnow team teams.

Overview

The first main pack is pro@coder, and like all future packs in this pro@ AI Pack namespace, they will be tailored toward Production Coding (versus Vibe Coding).

• Start a new project, but with production in mind from the start.

• Work on a big code base (10k lines or more) with the "globs lensing" parametric prompt (i.e., knowledge_globs and ``).

This is accomplished in a few ways:

• Interactive packs/agents will have a parametric prompt like pro@coder with the coder-prompt.md usually created in the project .aipack/.prompt/pro@coder/coder-prompt.md.
• These parametric prompt files allow the agent to let the user customize its behavior with some agent-specific parameters.
• For example, in pro@coder, the user can specify which model they want to use per request, focus the agent only on some of the files (with the context_globs), and even customize concurrency with working_globs and working_concurrency.
• pro@coder will update the files relative to the workspace when base_dir = .. is uncommented. Otherwise, its context will be the eventual answers below the ==== separator, and the globs won't be taken into account.

Licensing

All of these packs are licensed under MIT or Apache 2 for maximum flexibility and are, therefore, free to use as a base for creating your own.

Contribution

Except for finding obvious bugs or typos, it's recommended that you open an issue to describe the scope before working on and submitting a PR for a feature enhancement.

This GitHub Repo