hyperresearch
jordan-gibbs/hyperresearch
Hyperresearch transforms Claude Code into a comprehensive deep research agent executing a rigorous 16-step pipeline with persistent source vaults and adversarial review. It produces verified, fully cited reports ranging from quick briefs to full dissertations.
1.7k 191MITUpdated 2026-08-04
80
GreatOverview
Hyperresearch is an advanced deep research harness designed for Claude Code that orchestrates multi-step pipelines including width sweeps, contradiction graphs, depth investigations, and parallel adversarial critique. It features tier-adaptive execution modes, paywalled paper retrieval integration, tool-locked surgical patching to prevent unauthorized rewrites, and a persistent SQLite-backed knowledge vault for compounding insights across research sessions.
Capabilities
- ▸16-step adaptive research pipeline supporting light, full, and dissertation tiers
- ▸Parallel multi-perspective width sweeps fetching hundreds of sources per run
- ▸Automated paywall bypassing via Unpaywall and Europe PMC integration
- ▸Adversarial multi-critic review process with tool-locked surgical patching
- ▸Skeptical citation verification and hallucination auditing before report finalization
- ▸Persistent markdown and SQLite-indexed source vault for cross-session knowledge reuse
- ▸Configurable research profiles, gears, registers, and inference depths
Best for
Conducting multi-source academic and technical literature reviews with strict citation verification, Generating comprehensive dissertation chapters or lengthy analytical reports with full source provenance, Performing competitive landscape analyses and market surveys using parallel web fetching and independence auditing, Building and maintaining a persistent, searchable local knowledge vault of processed research sources
Works with
Claude Code
Evoa Score breakdown= Σ (score × weight)
80Task usefulness20%100 → +20.0
実タスクにどれだけ役立つか（機能の豊富さ・用途の明確さ）。 — AIによるcapabilities/use-cases解析
Code quality15%92 → +13.8
実装・指示の品質。 — AIによるSKILL.md/README解析
Maintenance15%100 → +15.0
リポジトリがどれだけ活発に保守されているか。 — GitHub 最終push日時の新しさ
Documentation12%54 → +6.5
ドキュメントの充実度・分かりやすさ。 — README/独自要約の情報量
Security15%50 → +7.5
危険・不審な挙動が無いか。 — AIによるセキュリティレビュー
Originality10%95 → +9.5
ありふれたラッパーではない独自性。 — AIによる独自性判定
Popularity8%83 → +6.6
コミュニティの採用度。 — GitHub Stars/Forks（対数スケール）
Compatibility5%25 → +1.3
対応AIエージェントの広さ。 — AIによる対応エージェント判定
Weighted total80.2 / 100
ライセンス不明/制限あり（Red）のSkillは総合スコアに0.85倍の補正を適用します。 ランキングはこのScoreのみで決まり、広告で変わりません。 算出方法の詳細 →
Security considerations
The skill executes local Python scripts and leverages subagents capable of fetching web content and interacting with local files. Users should be mindful of the URLs accessed and commands executed during deep research sweeps.
Categories
Summary and analysis are original content generated by AI Skills Rank. The skill's source text is not reproduced here — view it on the linked repository.