# Programmatic Search Agents: Extending Agentic Search Beyond Query Reformulation

arXiv: https://arxiv.org/abs/2610.06689 · HTML: https://mlsanity.com/paper/2610.06689/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Web, search, and research agents, one of 16 directions in LLM agents and tool use.
- Builds on 18 of its 34 references that are in the map.
- First posted on arXiv 5 Oct 2026.

## Placed in

- LLM agents and tool use / [Web, search, and research agents](https://mlsanity.com/agents/web-search-and-research-agents/)

## Builds on (most cited first, 18 of 18)

- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning](https://mlsanity.com/paper/2503.09516/) (2025)
- [BrowseComp: A Simple Yet Challenging Benchmark for Browsing Agents](https://mlsanity.com/paper/2504.12516/) (2025)
- [Reinforcement Learning via Self-Distillation](https://mlsanity.com/paper/2601.20802/) (2026)
- [ReWOO: Decoupling Reasoning from Observations for Efficient Augmented Language Models](https://mlsanity.com/paper/2305.18323/) (2023)
- [ReSearch: Learning to Reason with Search for LLMs via Reinforcement Learning](https://mlsanity.com/paper/2503.19470/) (2025)
- [Recursive Language Models](https://mlsanity.com/paper/2512.24601/) (2025)
- [AgentFold: Long-Horizon Web Agents with Proactive Context Management](https://mlsanity.com/paper/2510.24699/) (2025)
- [WebResearcher: Unleashing unbounded reasoning capability in Long-Horizon Agents](https://mlsanity.com/paper/2509.13309/) (2025)
- [An Empirical Study on Reinforcement Learning for Reasoning-Search Interleaved LLM Agents](https://mlsanity.com/paper/2505.15117/) (2025)
- [Lost in the Maze: Overcoming Context Limitations in Long-Horizon Agentic Search](https://mlsanity.com/paper/2510.18939/) (2025)
- [QUEST: Training Frontier Deep Research Agents with Fully Synthetic Tasks](https://mlsanity.com/paper/2605.24218/) (2026)
- [The Context Gathering Decision Process: A POMDP Framework for Agentic Search](https://mlsanity.com/paper/2605.07042/) (2026)
- [Argus: Evidence Assembly for Scalable Deep Research Agents](https://mlsanity.com/paper/2605.16217/) (2026)
- [One Policy, Any Budget: Internalizing Budget-Aware Search via Reinforcement Learning](https://mlsanity.com/paper/2609.00813/) (2026)
- [Delegation Intelligence in Deep Search: A Controllable Framework for Disentangled Capability Diagnosis](https://mlsanity.com/paper/2607.23524/) (2026)
- [LENS: In-Context Search via Latent Evidence Exploration over Dynamic Raw Documents](https://mlsanity.com/paper/2608.16185/) (2026)
- [Question's Gambit: The First Move Matters in Agentic Deep Search](https://mlsanity.com/paper/2609.14412/) (2026)

## Abstract (arXiv)

Search agents adapt their queries, yet fixed search interfaces leave candidate processing and evidence presentation outside the agent's direct control. Our trajectory analysis shows that supporting passages can be retrieved yet never delivered to the agent; a same-page oracle intervention shows that changing the returned evidence can reduce subsequent search. We introduce Programmatic Search Agent (PSA), which makes a local executable computation over candidates the unit of a search action. PSA unifies a persistent candidate workspace, flexible primitive composition, and selective evidence presentation. It incrementally generates program cells that reuse candidates, execute dependent operations, and select what the agent inspects next. The runtime resolves specified data dependencies within each cell, while the agent adapts its search strategy across cells as new evidence arrives. We compare PSA with the Query-based Agent and Tool-based Agent on InfoSeek-Eval and BrowseComp-Plus using five policy backbones without task-specific training. All three interfaces share the search substrate, and the Tool-based Agent also shares PSA's primitives and persistent workspace. Relative to the Query-based Agent, PSA improves macro-averaged task success by 4.00 and 7.56 percentage points on the two benchmarks, respectively; within-backbone reductions in final-step tokens average 28.3% and 33.9%. These results support extending agent control beyond query reformulation to the processing and presentation of retrieved evidence. Code will be released subject to approval.