AI Paper Atlas

Research map / Large language models / Agents, code and math

Just Ask: Curious Code Agents Reveal System Prompts in Frontier LLMs

Xiang Zheng, Yutao Wu, Hanxun Huang, Yige Li, Xingjun Ma, Bo Li, Yu-Gang Jiang, Cong Wang

ICML 2026 · Large language models › Agents, code and math · approach: Software engineering agents · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add UCB-guided skill selection instead of random selection (ablation)JustAsk · Framework ablation on Claude Sonnet 4, B=5Sim-GT: 0.757 -> 0.867Improved
add UCB-guided skill selection instead of random selection (ablation)JustAsk · Framework ablation on Claude Sonnet 4, B=10Sim-GT: 0.867 -> 0.889Improved
add skill framework to a bare agent (ablation)Bare-Agent · Framework ablation on Claude Sonnet 4, B=10Sim-GT: 0.801 -> 0.889Improved
increase query budget from 3 to 10 for bare agent (ablation)Bare-Agent · Framework ablation on Claude Sonnet 4Sim-GT: 0.801 -> 0.801No clear effect
use only introspection skill L14 instead of full skill set (ablation)JustAsk · Framework ablation on Claude Sonnet 4, B=5Sim-GT: 0.867 -> 0.771Hurt
add the skill framework to an unguided agent (Self-Evolving) (ablation)Unguided agent · End-to-end evaluation, 6 modelsSim-GT: +10.6% on average, improves 5 of 6 targetsImproved
add the skill framework to an unguided agent (Self-Evolving) (ablation)Unguided agent · End-to-end evaluation, Mistral LargeSim-GT: 0.55 -> 0.52 (-6.7%)Hurt
add the skill framework to an unguided agent (Self-Evolving) (ablation)Unguided agent · End-to-end evaluation, Llama 4 ScoutSim-GT: 0.73 -> 0.92 (+25.9%)Improved
add attack-aware defense in system promptFrontier models · Controlled defense evaluationSim-GT: 0.719 -> 0.587 (-18.4%)Improved
add simple do-not-reveal instructionFrontier models · Controlled defense evaluationSim-GT: 0.719 -> 0.676 (6.0% reduction)No clear effect
add simple do-not-reveal instructionClaude Opus 4.5 · Controlled defense evaluationSim-GT: 0.600 -> 0.616Hurt
use JustAsk instead of static baseline Raccoon on hardest targetRaccoon · Baseline comparison, B=3, target GPT-4oSim-GT: 0.701 -> 0.727Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy