code-stats
Analyze codebase with tokei (fast line counts by language) and difft (semantic AST-aware diffs). Get quick project overview without manual counting. Triggers on: how big is codebase, count lines of code, what languages, show semantic diff, compare files, code statistics.
96
Quality
100%
Does it follow best practices?
Impact
83%
3.95xAverage score across 3 eval scenarios
Passed
No known issues
Evaluation results
68%
38%Codebase Composition Report for Investor Due Diligence
tokei line count analysis
Uses tokei
0%
100%
No manual counting
100%
100%
Compact flag
0%
0%
Sort by code
0%
100%
Excludes node_modules
66%
100%
JSON output flag
0%
100%
jq for JSON extraction
0%
0%
Language type filter
0%
0%
Language breakdown present
100%
100%
Numeric extraction present
100%
100%
Without context: $0.3098 · 1m 24s · 23 turns · 22 in / 5,405 out tokens
With context: $0.5027 · 1m 34s · 28 turns · 27 in / 6,301 out tokens
100%
75%PR Review: "Cleanup" Commit Verification
difft semantic diff
Uses difft binary
0%
100%
Correct difft syntax
0%
100%
Inline display mode
0%
100%
Diff output captured
0%
100%
No traditional diff
0%
100%
Semantic change identified
100%
100%
Formatting noise dismissed
100%
100%
Without context: $0.2666 · 1m 21s · 12 turns · 16 in / 5,767 out tokens
With context: $0.5246 · 1m 52s · 29 turns · 32 in / 6,581 out tokens
81%
71%Release Notes: What Changed in the Latest Commit?
git semantic diff workflow
GIT_EXTERNAL_DIFF env var
0%
100%
git show with difft
0%
100%
git diff with difft
0%
0%
No plain git diff
0%
100%
Semantic output captured
0%
86%
Commands documented
0%
100%
No standalone difft
100%
50%
Without context: $0.7543 · 3m 3s · 39 turns · 42 in / 10,831 out tokens
With context: $0.5846 · 1m 55s · 34 turns · 38 in / 6,386 out tokens
- Repository
- NeverSight/skills_feed
- Commit
5342bca
- Evaluated
- Agent
- Claude Code
- Model
- Claude Sonnet 4.6
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.