deep-search-research-agent

Build research agents that systematically search for comprehensive answers to complex questions by maintaining search state, iterating on queries, and validating answer completeness. Implement strategies for identifying knowledge gaps and conducting follow-up searches to ensure thorough coverage of topics.

跳到安装

仓库: ADu2021/skillXiv
最近来源活动: 2026年3月24日 19:42
检测到的 SKILL.md 语言: 英语
星标: 6
分支: 0

安装方式

默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。

用 Codex 或 Claude 帮你安装复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。

直接命令不会经过审查 Prompt；运行前请先检查来源。

npx skills add https://github.com/ADu2021/skillXiv --skill deep-search-research-agent

命令会保持在同一行。复制前请横向滚动并检查完整内容。

想先保存到本地？可下载 SkillsMP 当前能够提供的文件。

检查来源文件

决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。

SKILL.md

来源说明 · 只读预览

返回搜索返回安装

meaningful-kebab-case-name

Convert arXiv papers into ready-to-use agent skills using category-aware extraction. First classifies the paper into one or more of 11 research categories, then applies a specialized extraction pipeline for each category — because different types of papers produce different types of usable knowledge. A single paper can yield multiple skills if it spans categories. Use this skill whenever the user wants to turn a paper into a skill, extract practical techniques from research, build a skill library from papers, convert arXiv papers into reusable agent instructions, or batch-process multiple papers into skills. Also trigger when someone asks about extracting actionable knowledge from papers, making research practical for LLM agents, or systematically converting academic contributions into structured agent capabilities.

2026-03-26

action-quantization-behavior-cloning

Establish regret bounds for behavior cloning with discretized actions combining statistical error and quantization error terms. Prove smoothness requirements for safe quantizer design, show that learning-based quantizers fail these requirements, and propose model-based augmentation to reduce error dependence from H² to H.

2026-03-26

adaptive-lora-personalized-ranks

Dynamically allocate LoRA ranks per-layer during fine-tuning instead of using fixed uniform ranks. Learn optimal rank for each layer and subject via variational framework with discretized exponential distribution, reducing memory footprint while maintaining fidelity and text-alignment.

2026-03-26

查看这个仓库的全部 Skills

class SearchState: """Maintain comprehensive search state""" def __init__(self, query): self.original_query = query self.search_history = [] self.discovered_answers = [] self.known_gaps = set() self.relevant_subtopics = set() self.coverage_map = {} (): search_record = { : query, : (results), : datetime.now(), : .summarize_results(results) } .search_history.append(search_record) result results: answer = .extract_answer(result) answer .discovered_answers: .discovered_answers.append(answer) (): .coverage_map = { : .covers_aspect(, .discovered_answers), : .covers_aspect(, .discovered_answers), : .covers_aspect(, .discovered_answers), : .covers_aspect(, .discovered_answers), : .covers_aspect(, .discovered_answers), : .covers_aspect(, .discovered_answers) } (): .update_coverage_map() gaps = [aspect aspect, covered .coverage_map.items() covered] .known_gaps = (gaps) gaps (): all_entities = .extract_all_entities(.discovered_answers) explored = {e e all_entities .has_sufficient_coverage(e)} all_entities - explored

class ComprehensivenessValidator: """Determine if search has yielded comprehensive answers""" def __init__(self, target_coverage=0.85): self.target_coverage = target_coverage def compute_coverage_score(self, state): """ Score how comprehensively the answer covers the question. Considers: breadth of dimensions, depth per dimension, diversity of sources """ state.update_coverage_map() # Dimension coverage: what % of question aspects are addressed dimensions = state.coverage_map.values() dimension_score = sum(dimensions) / len(dimensions) # Depth score: how many answers/perspectives per dimension answer_count = len(state.discovered_answers) depth_score = min(answer_count / 5, 1.0) # Normalize to 5+ answers # Diversity score: sources/perspectives vary source_diversity = self.measure_source_diversity(state.discovered_answers) diversity_score = source_diversity # Composite score: weighted combination coverage_score = ( 0.4 * dimension_score + 0.35 * depth_score + 0.25 * diversity_score ) return coverage_score def should_continue_searching(self, state, search_count): """ Determine if more searches are needed """ coverage = self.compute_coverage_score(state) gaps = state.identify_gaps() # Stop if coverage is high enough if coverage >= self.target_coverage: return False # Stop if too many searches already if search_count > 10: return False # Stop if diminishing returns (last 2 searches added <5% new info) if len(state.search_history) >= 3: recent_gain = self.compute_recent_information_gain(state, window=2) if recent_gain < 0.05: return False return True def compute_recent_information_gain(self, state, window=2): """Measure how much new information recent searches added""" if len(state.search_history) < window: return 1.0 recent_answers = set() for record in state.search_history[-window:]: recent_answers.update(record["results_summary"]) total_answers = set(state.discovered_answers) gain = len(recent_answers) / max(len(total_answers), 1) return gain

name	deep-search-research-agent
title	DeepSearchQA: Bridging the Comprehensiveness Gap for Deep Research Agents
version	0.0.2
engine	skillxiv-v0.0.2-claude-opus-4.6
license	MIT
url	https://arxiv.org/abs/2601.20975
keywords	["research-agent","information-retrieval","question-answering","deep-search","comprehensiveness"]
description	Build research agents that systematically search for comprehensive answers to complex questions by maintaining search state, iterating on queries, and validating answer completeness. Implement strategies for identifying knowledge gaps and conducting follow-up searches to ensure thorough coverage of topics.

name	deep-search-research-agent
title	DeepSearchQA: Bridging the Comprehensiveness Gap for Deep Research Agents
version	0.0.2
engine	skillxiv-v0.0.2-claude-opus-4.6
license	MIT
url	https://arxiv.org/abs/2601.20975
keywords	["research-agent","information-retrieval","question-answering","deep-search","comprehensiveness"]
description	Build research agents that systematically search for comprehensive answers to complex questions by maintaining search state, iterating on queries, and validating answer completeness. Implement strategies for identifying knowledge gaps and conducting follow-up searches to ensure thorough coverage of topics.

deep-search-research-agent

安装方式

检查来源文件

Problem

Solution

When to Use

When NOT to Use

Implementation

Step 1: Build Search State Management

Problem

Solution

When to Use

When NOT to Use

Implementation

Step 1: Build Search State Management

Step 2: Implement Gap-Driven Query Generation

Step 3: Build Comprehensiveness Validator

Step 4: Orchestrate the Deep Search Loop

Step 5: Answer Synthesis and Validation

Key Strategies

Success Indicators

References

deep-search-research-agent

来源信息

安装方式

检查来源文件

同仓库更多 Skills

同仓库更多 Skills

Problem

Solution

When to Use

When NOT to Use

Implementation

Step 1: Build Search State Management

Problem

Solution

When to Use

When NOT to Use

Implementation

Step 1: Build Search State Management

Step 2: Implement Gap-Driven Query Generation

Step 3: Build Comprehensiveness Validator

Step 4: Orchestrate the Deep Search Loop

Step 5: Answer Synthesis and Validation

Key Strategies

Success Indicators

References