rtk-ai/rtk

fix: ls tracking baseline should compare against `ls` not `ls -la`

Open

#561 opened on Mar 13, 2026

View on GitHub
 (1 comment) (4 reactions) (0 assignees)Rust (48,085 stars) (2,914 forks)batch import
bugeffort-smallfilter-qualitygood first issue

Description

Problem

rtk ls internally runs ls -la to get file metadata for grouping/filtering, then compresses the output. The savings % is tracked against the ls -la output, not plain ls (which is what the agent would have run without RTK).

This inflates the reported savings. Example from user report:

  • ls = 332 tokens
  • ls -la = 4173 tokens
  • rtk ls = 993 tokens
  • Reported: 76.2% savings (vs ls -la)
  • Actual: -199% vs plain ls

Fix

Track against the baseline command the agent would have run (ls), not the internal enriched command (ls -la).

In src/ls.rs line 113-118, change the tracking input from the ls -la raw output to a plain ls capture (or estimate).

Reference

https://www.reddit.com/r/ClaudeAI/comments/1rjkf9r/comment/oa3trnn/

Contributor guide

Tech stack
rust
Domain
cli
Issue type
bug
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
2
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
under 1 hour
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
fresh
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
clear
Prerequisites
basic Rust
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
80
Research direction
The issue is to fix the tracking baseline in `src/ls.rs` at lines 113-118. Currently, the savings are computed against `ls la` output, but should be against plain `ls`. The fix involves changing the input for tracking from the enriched `ls la` command to plain `ls`. The reported actual savings are negative compared to plain `ls`, so the calculation logic needs adjustment to capture the plain `ls` output (or estimate its token count). Refer to the Reddit thread linked for context.