Pith.

Software Engineering · arXiv:2603.15401 v2 · submitted 16 Mar 2026

SWE-Skills-Bench: Do Agent Skills Actually Help in Real-World Software Engineering?

Tingxu Han, Yi Zhang, Wei Song, Chunrong Fang, Zhenyu Chen, Youcheng Sun, Lijie Hu

The paper says

“Our results show that skills provide limited benefits: 92 of 98 skills have no statistically significant effect, and adding skills raises the average pass ratio only from 76.6% to 76.9%.”

from the abstract · picked out as the main claim by a classifier

Pith says

Pith has not reviewed this paper.

Get a Pith Review, $5

Reviews

Reviews by people

No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.

Request a human review

How Pith checks papers · The full audit of this paper

The paper

Abstract

Agent skills, structured procedural knowledge packages injected at inference time, are increasingly used to augment LLM agents on software engineering tasks. However, their real utility in end-to-end development settings remains unclear, and no standard procedure tests an individual skill before installation. We present swe-skills-bench, a skill-centric, requirement-driven benchmark that measures how much individual agent skills improve performance on real-world software engineering (SWE) tasks. We start from 98 public SWE skills, each paired with a real GitHub repository pinned at a fixed commit. For each skill, we derive repository-specific requirements, create executable verifiers, and build a reproducible environment. Under each of four agent-model configurations, we run every task once without the skill and once with it, yielding 3,920 paired evaluations. For each requirement, we compute a paired delta, i.e., the change in outcome when the skill is added. We then aggregate these deltas across a skill's ten requirements into a per-skill utility profile, and use a statistical test to classify each skill as Golden (significantly helpful), Ineffective (no significant effect), or Harmful (significantly detrimental). Our results show that skills provide limited benefits: 92 of 98 skills have no statistically significant effect, and adding skills raises the average pass ratio only from 76.6% to 76.9%. The few helpful skills supply procedures that the agent would otherwise omit (gains of up to + 17.7 points). In contrast, harmful skills anchor the agent on their own templates and override the task's explicit instructions (losses of up to -17.0 points). Our code is available at https://github.com/ControllableModel/SWE-Skills-Bench.

The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. No sentence in this record was written by a model.

abstract compiled · 2026-10-28 · no language-model review

How the main claim is stated: asserted 1.00

What the abstract itself states, as probabilities: claims first 0.29 · states limitation 0.50 · human participants 0.03 · controlled comparison 0.97 · randomized assignment 0.03 · states quantitative result 0.99 · compares with prior results 0.64 · mentions code or data availability 0.98

The abstract, sentence by sentence

This table labels the abstract only (arXiv listing abstract v2). The number is the probability the classifier gave the label.

SentenceRoleThe authors' own workStates the main conclusionStrength
Agent skills, structured procedural knowledge packages injected at inference time, are increasingly used to augment LLM agents on software engineering tasks. background 1.00 0.04 0.02 asserted 0.98
However, their real utility in end-to-end development settings remains unclear, and no standard procedure tests an individual skill before installation. gap 1.00 0.05 0.03 asserted 0.89
We present swe-skills-bench, a skill-centric, requirement-driven benchmark that measures how much individual agent skills improve performance on real-world software engineering (SWE) tasks. contribution 1.00 0.95 0.06 asserted 1.00
We start from 98 public SWE skills, each paired with a real GitHub repository pinned at a fixed commit. method 0.95 0.89 0.02 asserted 0.99
For each skill, we derive repository-specific requirements, create executable verifiers, and build a reproducible environment. method 1.00 0.95 0.02 asserted 0.98
Under each of four agent-model configurations, we run every task once without the skill and once with it, yielding 3,920 paired evaluations. method 1.00 0.97 0.02 asserted 0.98
For each requirement, we compute a paired delta, i.e., the change in outcome when the skill is added. method 1.00 0.96 0.02 asserted 0.98
We then aggregate these deltas across a skill's ten requirements into a per-skill utility profile, and use a statistical test to classify each skill as Golden (significantly helpful), Ineffective (no significant effect), or Harmful (significantly detrimental). method 1.00 0.96 0.05 asserted 0.95
Our results show that skills provide limited benefits: 92 of 98 skills have no statistically significant effect, and adding skills raises the average pass ratio only from 76.6% to 76.9%. result 1.00 0.97 0.88 asserted 1.00
The few helpful skills supply procedures that the agent would otherwise omit (gains of up to $+17.7$ points). result 0.81 0.95 0.11 asserted 1.00
In contrast, harmful skills anchor the agent on their own templates and override the task's explicit instructions (losses of up to -17.0 points). result 0.59 0.95 0.10 asserted 1.00
Our code is available at https://github.com/ControllableModel/SWE-Skills-Bench. contribution 0.88 0.39 0.02 asserted 0.93

What kind of paper the classifier reports this is: dataset or benchmark 0.96 · evidence from observation or experiment 0.99

What was examined, and what was not

Sources read

  • arXiv listing abstract v2

Not examined

  • full text
  • figures and tables
  • references and citing sentences
  • relations to other papers

Levels run

  • P abstract

Not run

  • S sections
  • T body sentences
  • C citations
  • R cross-paper relations

How the labels were made

Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-10-28: 1 request, 8,946 input tokens.

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. This record as JSON

Cited by 7 Pith papers
How the literature uses it

48 arXiv papers cite this in 81 sentences

Each citing sentence is taken from the citing paper's LaTeX source and typed by Pith's classifier. A label appears at probability 0.8 or more (0.95 for support), a contest at 0.7 or more; the quoted words are the citing authors'. How each label was checked · JSON

Papers citing it · 50 sentences without a confident label

Uses its method or tool 1 paper
  • Following [ref] on evaluating coding-agent performance, we measure and report: Wrap success ($↑$) measures whether the agent's generated integration loads the released checkpoint and completes the benchmark evaluation protocol end-to-end without manual intervention.

    Nautilus: From One Prompt to Plug-and-Play Robot Learning · Ablation: from pairwise wiring to reusable extending · 0.88

Uses its data or values 2 papers
Cites it as support 5 papers
Cites it as background 18 papers

Challenges and discussion

Know something about this paper? Leave a comment.

No comments yet.

Sign in to comment