Metadata-Version: 2.4
Name: knackpack
Version: 1.1.0rc1
Summary: Knackpack — Engineering Context Platform for AI coding assistants (Builder CLI)
Project-URL: Homepage, https://knackpack.co.in
Project-URL: Documentation, https://docs.knackpack.co.in
Author: Knackpack
License: Proprietary
License-File: LICENSE
Keywords: ai,coding-assistant,cursor,developer-tools,engineering-context
Classifier: Development Status :: 5 - Production/Stable
Classifier: Environment :: Console
Classifier: Intended Audience :: Developers
Classifier: License :: Other/Proprietary License
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Topic :: Software Development
Requires-Python: >=3.12
Requires-Dist: httpx>=0.27
Requires-Dist: packaging>=24.0
Requires-Dist: pydantic>=2.9
Requires-Dist: pyyaml>=6.0.2
Requires-Dist: questionary>=2.0
Requires-Dist: rich>=13.7
Requires-Dist: typer>=0.27.0
Description-Content-Type: text/markdown

# Knackpack Builder

**Knackpack is an Engineering Context Platform for AI coding assistants.**
It makes assistants like Cursor understand your project the way an
experienced software engineer does — the stack, the architecture, the
dependency graph, the decisions, and the risks — all generated locally, with
evidence, in files your team can read and commit.

Builder Packs give your assistant professional skills; the Knackpack CLI links
them into a project and generates your Engineering Context. Pack skill content
is fetched live from the Knackpack Engine via `knackpack brief` — it is not
stored in your repo.

## Get started

```sh
uv tool install knackpack

cd my-project
knackpack login                              # approve the device in your browser
knackpack install founding-builder-pack
knackpack init                               # generate the Engineering Context
```

Then open the project in Cursor — its pack rule selects and loads only the
knack relevant to each task. Full walkthrough: [docs/quickstart.md](docs/quickstart.md).

## What `knackpack init` gives your assistant

| Artifact | What it contains |
| --- | --- |
| `context/project.json` | Detected stack: languages, frameworks, databases, auth, package managers, deployment, AI SDKs — every fact with file evidence |
| `context/architecture.md` | Architecture summary (yours to extend below the markers) |
| `context/decisions/` | Your engineering decisions, indexed but never rewritten |
| `graph/graph.json` | Modules, entry points, internal + external dependency edges |
| `context/future-warnings.md` | Evidence-backed risks, before they bite |

Everything is local. No LLM calls, no code leaves your machine, telemetry is
a no-op.

## Documentation

- [Installation](docs/installation.md) · [Quick Start](docs/quickstart.md)
- [Command Reference](docs/commands.md) · [Troubleshooting](docs/troubleshooting.md)
- [FAQ](docs/faq.md) · [Best Practices](docs/best-practices.md)
- [Changelog](CHANGELOG.md)

## Development

Requires [uv](https://docs.astral.sh/uv/); Python 3.12+ is provisioned
automatically.

```sh
make install   # sync the environment
make check     # ruff + mypy --strict + import-linter + pytest
```

Architecture rules live in [`CLAUDE.md`](CLAUDE.md) and are enforced
mechanically by import-linter. Design decisions are ADRs in
[`docs/adr/`](docs/adr/); milestone reports in
[`docs/milestones/`](docs/milestones/). The mock registry under
`tests/support/mock_registry/` is the executable specification of the
registry `/v1` API.
