Read original ↗
paperarXivTrust 82 · PrimaryPublished 1mo agoLive · 1mo ago

SWE-Doctor: Guiding Software Engineering Agents with Runtime Diagnosis from Multi-Faceted Bug Reproduction Tests

Large language model (LLM)-based software engineering agents are increasingly developed to resolve software issues by generating patches from issue reports and code repositories. Bug reproduction tests (BRTs) are an important building block for such agents and have been shown useful for patch validation. However, it remains unclear whether BRTs can also help the more central stage of patch generation. We first conduct a preliminary study and find that directly using advanced BRT generators to guide patch generation is not beneficial: fail-to-fail BRTs can mislead agents, while even fail-to-pas

Lineage graph

Paper → model → repo connections mined from source citations (Tier-1 exact match).

Why these links exist

Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.

Covers

newsA system-level approach to prompt injection: separating instruction and data channels in LLM agents [P]newsPatch the Planet: a Daybreak initiative to support open source maintainersnewsDebugging production agents with Amazon Bedrock AgentCore Observability

Implements

repoagent-tools

Related to

toolAgentTrace

Implements (incoming)

repobug-ops/zephrepomozilla/bugbugrepopromptfoo/promptfoorepogolobokov.misha/llm-review-agentsrepohuhusmang/Awesome-LLMs-for-Vulnerability-DetectionrepoNayjest/Gitoreporaphaelmansuy/code2promptrepoFailproofAI/failproofairepogigioneggiando/argo

Covers (incoming)

newsMicrosoft patches record number of security vulnerabilities, citing its use of AInewsCode fixers have fired up the AI warp drive. Strange new worlds await

Related across the graph

repobug-ops/zephrepomozilla/bugbugreporaphaelmansuy/code2promptnewsDebugging production agents with Amazon Bedrock AgentCore ObservabilityrepoNayjest/GitonewsCode fixers have fired up the AI warp drive. Strange new worlds awaitrepogigioneggiando/argonewsPatch the Planet: a Daybreak initiative to support open source maintainersnewsMicrosoft patches record number of security vulnerabilities, citing its use of AIrepohuhusmang/Awesome-LLMs-for-Vulnerability-DetectionrepoFailproofAI/failproofainewsA system-level approach to prompt injection: separating instruction and data channels in LLM agents [P]repopromptfoo/promptfoorepogolobokov.misha/llm-review-agentsrepoagent-toolstoolAgentTrace

Topics

cs.AI