noslop

Reviews branch diffs to delete low-value tests and filler prose.

3.2k|266|Updated Mar 22, 2024
One-click install
npx skills add https://github.com/marin-community/marin --skill noslop
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: noslop
Source: https://github.com/marin-community/marin/tree/main/.agents/skills/noslop
Command: npx skills add https://github.com/marin-community/marin --skill noslop

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

Code review diffs often accumulate low-value tests that assert implementation details and prose filled with stock phrases, empty framing, and unsupported claims. This Skill systematically identifies and removes that slop from a branch diff before merge.

Core Features & Use Cases

  • Behavioral Test Gating: Evaluates every changed test against whether it fails when real user-visible behavior breaks, deleting tests that only check constructor assignment, types, log text, or mock call counts.
  • Prose Slop Detection: Scans comments, docstrings, docs, commit messages, and PR text for stock contrasts, bridge phrases, AI vocabulary, and unsupported claims using a pattern-scanning script.
  • Design-Level Cleanup: Reviews new files, helpers, flags, and compatibility paths in the diff, removing code whose consumers disappeared.
  • Use Case: Before merging a feature branch, run the review to delete a test that only asserts isinstance on a config object and rewrite a PR description claiming unmeasured speedups.

Quick Start

Use the noslop skill to review this branch diff against main and remove low-value tests, filler prose, and leftover complexity.

Frequently Asked Questions about noslop

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I review a git diff for low-value tests?

Compute the merge-base diff against main, then check each changed test against whether it fails when real user-visible behavior breaks. Delete tests asserting implementation details like constructor assignment, isinstance checks, mock call counts, or exact log text.

What test patterns should be deleted in code review?

Delete tests asserting config round-trips, type checks, private state, assert_called_once_with, exact log prose, is-not-None checks, and golden outputs generated by the implementation under test. Keep tests protecting real boundaries, regressions, invariants, and wire contracts.

How do I detect AI-generated filler phrases in documentation?

Scan added lines for stock contrasts like 'not X, but Y', bridge phrases like 'it is worth noting' and 'stepping back', and vocabulary like 'crucial', 'seamless', and 'delve'. The scan_diff.py script automates this pattern matching over the diff.

Does the diff scanner include untracked files?

Yes, the scan_diff.py script includes untracked files via git ls-files in addition to added diff lines. It filters them by prose file suffixes and test path patterns before applying the pattern catalogs.

When should a weak test be kept instead of deleted?

Keep a test when it protects a real boundary, regression, invariant, round-trip, state transition, persisted effect, wire contract, or numerical result against an independent reference. Exact prose assertions are acceptable only when another program parses that text.