WebChallenger introduces PageMem and three architecture mechanisms to achieve competitive web navigation with open-weight LLMs on WebArena, VisualWebArena, Online-Mind2Web, and WorkArena without fine-tuning or site adapters.
hub
Inducing programmatic skills for agentic tasks
21 Pith papers cite this work. Polarity classification is still indexing.
21
Pith papers citing it
abstract
To succeed in common digital tasks such as web navigation, agents must carry out a variety of specialized tasks such as searching for products or planning a travel route. To tackle these tasks, agents can bootstrap themselves by learning task-specific skills online through interaction with the web environment. In this work, we demonstrate that programs are an effective representation for skills. We propose agent skill induction (ASI), which allows agents to adapt themselves by inducing, verifying, and utilizing program-based skills on the fly. We start with an evaluation on the WebArena agent benchmark and show that ASI outperforms the static baseline agent and its text-skill counterpart by 23.5% and 11.3% in success rate, mainly thanks to the programmatic verification guarantee during the induction phase. ASI also improves efficiency by reducing 10.7-15.3% of the steps over baselines, by composing primitive actions (e.g., click) into higher-level skills (e.g., search product). We then highlight the efficacy of ASI in remaining efficient and accurate under scaled-up web activities. Finally, we examine the generalizability of induced skills when transferring between websites, and find that ASI can effectively reuse common skills, while also updating incompatible skills to versatile website changes.
hub tools
citation-role summary
background 4
citation-polarity summary
roles
background 3polarities
background 3representative citing papers
Framework estimates context-dependent marginal utility of candidate skills via reward gaps in matched base vs. skill-augmented rollouts to filter skills and co-train policy as generator.
SelSkill applies dual-granularity preference learning to selective skill-or-skip decisions, improving task success by 10.9 points and execution precision by 29.1 points on ALFWorld with Qwen3-8B.
SkillSafetyBench is a benchmark of 155 cases across 47 tasks and 6 risk domains showing that non-user attacks via skills, artifacts, or environments can consistently induce unsafe agent behavior.
OLIVIA treats LLM agent action selection as a contextual linear bandit over frozen hidden states and applies UCB exploration to adapt online, yielding consistent gains over static ReAct and prompt-based baselines on four benchmarks.
This survey introduces the Generate-Filter-Control-Replay (GFCR) taxonomy to structure rollout pipelines for RL-based post-training of reasoning LLMs.
MAS-Bench introduces 139 tasks, 88 predefined shortcuts, and 9 metrics to evaluate hybrid GUI-shortcut mobile agents, reporting up to 68.3% success and 39% efficiency gains over GUI-only baselines.
Multimodal skills retaining visual figures improve CUA benchmark scores by 8.3 points over text-only equivalents generated from the same source content.
SkillMigrator reduces LLM-action counts by 8-10% on WebArena and Mind2Web by transferring web skills via layout-matched transferable interaction patterns.
Catalogs ten patterns and synthesizes a four-layer reference architecture for skill harnessing in LLM agents, evaluated via cross-instantiation on eight systems.
An iterative framework lets LLMs learn procedural assessment skills for rubric construction, improving automated scoring on all ten ASAP-SAS items and often exceeding expert rubrics while showing cross-item transfer.
Maestro uses outcome-based RL to train a lightweight policy that orchestrates ensembles of frozen expert models and skills, reporting 70.1% average accuracy across ten multimodal benchmarks and outperforming GPT-5 and Gemini-2.5-Pro while generalizing to unseen components.
SkillGraph jointly evolves agent skills and collaboration topologies in multi-agent vision-language systems using a multimodal graph transformer and a skill designer, yielding consistent performance gains on benchmarks.
SkillTrojan backdoors skill-based agents by partitioning an encrypted payload across benign-looking skills that reassemble and execute only under a predefined trigger.
CoM organizes memory fragments into evolving inference paths with adaptive truncation, delivering 7.5-10.4% accuracy gains on long-memory benchmarks at 2.7% token cost and 6% latency of complex alternatives.
EvoSOP enables LLM agents to self-evolve by iteratively synthesizing atomic tool actions into reusable Standard Operating Procedures, improving success rates and reducing reasoning rounds on ACEBench and Tau2Bench.
SkillDisCo distills reusable PFSM subgraphs from successful agent traces and compiles them into callable procedural skills, improving success rates and reducing turns on ALFWorld and WebArena.
This survey categorizes agentic environments for LLMs by eight attributes and domains, introduces symbolic and neural synthesis paradigms with evaluation, and outlines four agent evolution pathways plus three environment evolution paradigms.
DataCOPE uses verifier-guided contrastive distillation from agent trajectories to discover skills, yielding average gains of 9.71% on report-style and 32.30% on reasoning-style data analysis tasks across four model settings.
LLM agent progress depends on externalizing cognitive functions into memory, skills, protocols, and harness engineering that coordinates them reliably.
A survey that defines agent skills as reusable procedural artifacts and reviews methods, resources, and applications across their representation, acquisition, retrieval, and evolution stages.
citing papers explorer
-
WebChallenger: A Reliable and Efficient Generalist Web Agent
WebChallenger introduces PageMem and three architecture mechanisms to achieve competitive web navigation with open-weight LLMs on WebArena, VisualWebArena, Online-Mind2Web, and WorkArena without fine-tuning or site adapters.
-
Co-Evolving Skill Generation and Policy Optimization
Framework estimates context-dependent marginal utility of candidate skills via reward gaps in matched base vs. skill-augmented rollouts to filter skills and co-train policy as generator.
-
Skill or Skip? Learning Selective Skill Invocation in Agentic Tasks via Dual-Granularity Preference Learning
SelSkill applies dual-granularity preference learning to selective skill-or-skip decisions, improving task success by 10.9 points and execution precision by 29.1 points on ALFWorld with Qwen3-8B.
-
SkillSafetyBench: Evaluating Agent Safety under Skill-Facing Attack Surfaces
SkillSafetyBench is a benchmark of 155 cases across 47 tasks and 6 risk domains showing that non-user attacks via skills, artifacts, or environments can consistently induce unsafe agent behavior.
-
OLIVIA: Online Learning via Inference-time Action Adaptation for Decision Making in LLM ReAct Agents
OLIVIA treats LLM agent action selection as a contextual linear bandit over frozen hidden states and applies UCB exploration to adapt online, yielding consistent gains over static ReAct and prompt-based baselines on four benchmarks.
-
Generate, Filter, Control, Replay: A Comprehensive Survey of Rollout Strategies for LLM Reinforcement Learning
This survey introduces the Generate-Filter-Control-Replay (GFCR) taxonomy to structure rollout pipelines for RL-based post-training of reasoning LLMs.
-
MAS-Bench: A Unified Benchmark for Shortcut-Augmented Hybrid Mobile GUI Agents
MAS-Bench introduces 139 tasks, 88 predefined shortcuts, and 9 metrics to evaluate hybrid GUI-shortcut mobile agents, reporting up to 68.3% success and 39% efficiency gains over GUI-only baselines.
-
VISUALSKILL: Multimodal Skills for Computer-Use Agents
Multimodal skills retaining visual figures improve CUA benchmark scores by 8.3 points over text-only equivalents generated from the same source content.
-
Beyond Domains: Reusing Web Skills via Transferable Interaction Patterns
SkillMigrator reduces LLM-action counts by 8-10% on WebArena and Mind2Web by transferring web skills via layout-matched transferable interaction patterns.
-
Harnessing Agent Skills: Architectural Patterns and a Reference Architecture for Skill-Mediated LLM Agents
Catalogs ten patterns and synthesizes a four-layer reference architecture for skill harnessing in LLM agents, evaluated via cross-instantiation on eight systems.
-
Learnable Assessment Skills for LLM-based Automated Scoring: Rubric Construction via Iterative Optimization
An iterative framework lets LLMs learn procedural assessment skills for rubric construction, improving automated scoring on all ten ASAP-SAS items and often exceeding expert rubrics while showing cross-item transfer.
-
Maestro: Reinforcement Learning to Orchestrate Hierarchical Model-Skill Ensembles
Maestro uses outcome-based RL to train a lightweight policy that orchestrates ensembles of frozen expert models and skills, reporting 70.1% average accuracy across ten multimodal benchmarks and outperforming GPT-5 and Gemini-2.5-Pro while generalizing to unseen components.
-
SkillGraph: Self-Evolving Multi-Agent Collaboration with Multimodal Graph Topology
SkillGraph jointly evolves agent skills and collaboration topologies in multi-agent vision-language systems using a multimodal graph transformer and a skill designer, yielding consistent performance gains on benchmarks.
-
SkillTrojan: Backdoor Attacks on Skill-Based Agent Systems
SkillTrojan backdoors skill-based agents by partitioning an encrypted payload across benign-looking skills that reassemble and execute only under a predefined trigger.
-
Chain-of-Memory: Lightweight Memory Construction with Dynamic Evolution for LLM Agents
CoM organizes memory fragments into evolving inference paths with adaptive truncation, delivering 7.5-10.4% accuracy gains on long-memory benchmarks at 2.7% token cost and 6% latency of complex alternatives.
-
From Atomic Actions to Standard Operating Procedures: Iterative Tool Optimization for Self-Evolving LLM Agents
EvoSOP enables LLM agents to self-evolve by iteratively synthesizing atomic tool actions into reusable Standard Operating Procedures, improving success rates and reducing reasoning rounds on ACEBench and Tau2Bench.
-
SKILL-DISCO: Distilling and Compiling Agent Traces into Reusable Procedural Skills
SkillDisCo distills reusable PFSM subgraphs from successful agent traces and compiles them into callable procedural skills, improving success rates and reducing turns on ALFWorld and WebArena.
-
Agentic Environment Engineering for Large Language Models: A Survey of Environment Modeling, Synthesis, Evaluation, and Application
This survey categorizes agentic environments for LLMs by eight attributes and domains, introduces symbolic and neural synthesis paradigms with evaluation, and outlines four agent evolution pathways plus three environment evolution paradigms.
-
Unsupervised Skill Discovery for Agentic Data Analysis
DataCOPE uses verifier-guided contrastive distillation from agent trajectories to discover skills, yielding average gains of 9.71% on report-style and 32.30% on reasoning-style data analysis tasks across four model settings.
-
Externalization in LLM Agents: A Unified Review of Memory, Skills, Protocols and Harness Engineering
LLM agent progress depends on externalizing cognitive functions into memory, skills, protocols, and harness engineering that coordinates them reliably.
-
A Comprehensive Survey on Agent Skills: Taxonomy, Techniques, and Applications
A survey that defines agent skills as reusable procedural artifacts and reviews methods, resources, and applications across their representation, acquisition, retrieval, and evolution stages.