rtk-ai/rtk

Feature Request: rtk uv Support

Open

#2080 aperta il 24 mag 2026

Vedi su GitHub
 (1 commento) (6 reazioni) (0 assegnatari)Rust (2914 fork)batch import
area:clienhancementhelp wantedpriority:medium

Metriche repository

Star
 (48.085 star)
Metriche merge PR
 (Merge medio 11g 1h) (45 PR mergiate in 30 g)

Descrizione

Summary

Add generic support for:

rtk uv <command>

allowing RTK to work as a transparent proxy/filter layer on top of uv commands.

Examples:

rtk uv run pytest
rtk uv run ruff check .
rtk uv sync
rtk uv add requests

Motivation

Today, Python projects managed with uv require users and agents to execute commands directly with:

uv <command>

which bypasses RTK’s filtering and output-cleanup layer.

The proposal is to allow RTK to generically wrap uv commands instead of implementing isolated integrations for individual tools.

This would let RTK operate naturally in uv-managed environments while preserving its philosophy around:

  • transparent execution
  • command rewriting
  • noise reduction
  • workflow preservation
  • AI-friendly terminal output

Proposed Behavior

Instead of:

uv <command>

users could execute:

rtk uv <command>

with RTK acting as the execution/filter layer on top of uv.

Examples:

rtk uv run pytest
rtk uv run ruff check .
rtk uv sync
rtk uv add requests

Reuse Existing RTK Filters

One important aspect of this proposal is that the generic uv support could reuse RTK’s existing command filters and cleanup logic.

For example:

rtk uv run pytest

could internally reuse the same filtering/output-processing pipeline already used by:

rtk pytest

The same idea applies to other tools already supported by RTK.

This would avoid creating duplicated integrations while keeping behavior consistent across direct and uv-managed executions.


Additional Benefit: Cleaner Installation Output

Another major advantage would be reducing unnecessary installation/bootstrap logs generated by uv.

Current output frequently includes verbose setup noise such as:

Resolved 143 packages in 8ms
Audited 143 packages in 4ms
Prepared 2 packages
Installed 2 packages

RTK could preserve:

  • dependency resolution failures
  • installation errors
  • important warnings

while filtering non-essential success logs and repetitive bootstrap output.

This would significantly improve readability for both developers and AI agents.


Suggested Architecture

Potential execution flow:

rtk uv <command>
        ↓
RTK intercept layer
        ↓
Execute uv <command>
        ↓
Reuse existing RTK command filters
        ↓
Suppress installation/bootstrap noise
        ↓
Return compact output

Why Generic Support Matters

The proposal is intentionally generic and not tied to specific tools like pytest or ruff.

The goal is to support:

rtk uv <any-command>

so RTK can naturally integrate with modern Python workflows powered by uv.

This keeps the solution scalable and aligned with RTK’s architecture as a transparent execution and filtering layer.


Expected Outcome

The goal is to make RTK work seamlessly with uv while preserving RTK’s core value proposition:

  • cleaner outputs
  • reduced terminal noise
  • improved AI-agent context
  • transparent developer experience
  • minimal workflow changes

Related to #1205, #1379, #1405, #1641

Guida contributor