Fetching from the wire…
01
02
03
04
05
06
07
08
Research2026-08-21 · source-backed
119 repository-level tasks from 98 GitHub repos across 20 scientific domains, split into issue-driven, expert-exploratory, and engineering-integration paradigms. Claude Code with Opus-5 (max) lands below 50%. arXiv The ablation is the better finding: stripping explicit scientific guidance while keeping repo and execution context shows domain knowledge isn't uniformly helpful. Well-grounded facts improve average performance and token efficiency; poorly aligned guidance anchors the agent and doesn't raise exact repair success. That's the same anchoring mechanism MemTrapBench found in memory, showing up in a completely different input channel.
Each link below shares sources, entities, or timing with this story.
Claude Code uses Opus / Shared entities / Shared topic / Earlier coverage / Tension
Linked by a graph relationship (Claude Code uses Opus); both cover Claude Code, GitHub, Opus; overlapping topics (agent, best, claude, code, repo).
Linked by a graph relationship (Claude Code uses Opus); both cover Claude Code, GitHub, Opus, SWE; overlapping topics (claude, code).
Claude Code uses MCP / Shared entities / Shared topic / Earlier coverage
Linked by a graph relationship (Claude Code uses MCP); both cover Claude Code, GitHub, Opus, SWE; overlapping topics (agent, claude, code).
Graft supports Claude Code / Shared entities / Shared topic / Earlier coverage
Linked by a graph relationship (Graft supports Claude Code); both cover Claude Code, GitHub, SWE; overlapping topics (agent, claude, code, context, repo).
Anthropic released Claude Code / Shared entities / Same source domain / Shared topic / Earlier coverage
Linked by a graph relationship (Anthropic released Claude Code); both cover Claude Code, GitHub, SWE; reported by the same outlet (arxiv.org).
AgentRadio uses Claude Code / Shared entities / Same source domain / Shared topic / Earlier coverage
Linked by a graph relationship (AgentRadio uses Claude Code); both cover Claude Code, Opus, SWE; reported by the same outlet (arxiv.org).
Claude Code uses Haiku / Shared entities / Same source domain / Shared topic / Earlier coverage
Linked by a graph relationship (Claude Code uses Haiku); both cover Claude Code, Opus; reported by the same outlet (arxiv.org).
Codex competes with Claude Code / Shared entities / Same source domain / Shared topic / Earlier coverage
Linked by a graph relationship (Codex competes with Claude Code); both cover Claude Code, SWE; reported by the same outlet (arxiv.org).