Worker for analysing one academic text — a book chapter, a journal paper, or a talk transcript. Writes exactly one structured markdown output.
Worker for running quasi-audit on a file or directory, applying only local minimal fixes, and returning structured audit_result.
Worker for offline citation context-fit checks. Reads draft context and vault candidates, then writes review cards for one batch.
Worker for acquiring academic source files. Fetches, inspects, accepts matching files, and returns per-item status.
Worker for extracting chapter text from one EPUB/PDF book source. Writes a chapter manifest and text files.
Use when the user wants to build an author profile from a scholar's representative books and papers.
Use when the user wants to process a book from title, author, ISBN, or source file into chapter analyses and a book overview.
Use when the user wants to finalise a draft by proofreading text and checking citations, or clean up proofread records after review.
Use when the user wants to search, download, and analyse a single academic paper from a DOI, title, author, or existing source file.
Use when the user wants to transcribe and summarise a meeting or lecture recording into a structured talk page in the vault.
Executes bash commands
Hook triggers when Bash tool is used
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
This plugin requires configuration values that are prompted when the plugin is enabled. Sensitive values are stored in your system keychain.
soniox_api_keyAPI key for the Soniox stt-async-v4 speech-to-text engine, used by process-talk as the highest-quality transcription engine in the ensemble. Get one at https://soniox.com. Stored in the system keychain.
${user_config.soniox_api_key}superset_agentAgent name passed to superset agents create --agent for delegated quasi workflows. Leave empty to use copilot.
${user_config.superset_agent}anna_donator_keyDonator key from https://annas-archive.org/donations — unlocks the Fast Download API used for batch book acquisition. Stored in the system keychain.
${user_config.anna_donator_key}cookiecloud_uuidThe 'user KEY' set in the CookieCloud Chrome extension. Stored in the system keychain.
${user_config.cookiecloud_uuid}cookiecloud_serverEndpoint of your CookieCloud server used to pull browser cookies for EZProxy downloads, e.g. https://cookiecloud.example.com. Leave empty to disable EZProxy entirely. Stored in the system keychain.
${user_config.cookiecloud_server}immersive_auth_keyAuth key for the Immersive Translate Zotero PDF translation API. Used by the translate skill. Stored in the system keychain.
${user_config.immersive_auth_key}kagi_session_tokenKagi Session Link URL or raw session token for Douban Chinese-edition discovery. Stored in the system keychain and passed to kagi-cli as KAGI_SESSION_TOKEN.
${user_config.kagi_session_token}cookiecloud_passwordEnd-to-end encryption password set in the CookieCloud Chrome extension. Stored in the system keychain.
${user_config.cookiecloud_password}cookiecloud_ezproxy_domainCookie domain used by your institution's EZProxy after login, e.g. .idm.oclc.org or ezproxy.example.edu. This is not the login URL; only cookies on this exact domain are kept. Stored in the system keychain.
${user_config.cookiecloud_ezproxy_domain}cookiecloud_ezproxy_base_urlBase URL of your institution's EZProxy, e.g. https://ezproxy.example.edu. quasi appends /login?url= internally. Stored in the system keychain.
${user_config.cookiecloud_ezproxy_base_url}仿佛读过、仿佛想过、仿佛写过。
Claude Code 知识库插件。quasi 的维护逻辑是: 上层只看
skills → agents → bin/quasi-*,底层实现收束在少数大入口脚本里,方便
agent 一次读完整条能力链。
skills/ # 用户工作流编排
agents/ # LLM 代理壳,只调用 quasi-* CLI
bin/ # 稳定外部入口
scripts/ # deterministic 能力入口和实现
scripts/schemas/ # vault 领域规范(Pydantic + body schema)
core/ # 极小运行时地基(path/frontmatter/json/module loading)
scripts/schemas/ 不是 agent-facing API。agent 只依赖 CLI; schema 只给
typecheck、audit、citation biblio、migration 等 deterministic scripts 使用。
Skill 写作 schema 的维护者约定见 docs/SKILL_ORCHESTRATION.md:skill 主进程
owns state,agent 只做专业工种,每个 phase 必须有明确的 skip/failure/human gate。
active skill 正文只保留运行时需要的信息。
| Skill | 功能 |
|---|---|
process-book | 自动获取图书,提取章节,逐章分析并生成全书综述 |
process-paper | 搜索/下载/分析单篇论文 |
process-author | 获取代表作品并生成学者档案 |
process-topic | 主题语料处理 |
process-talk | 录制转写(多引擎集成)+ 结构化摘要入库 vault/talks/ |
wrap-up | draft 校对 + 引文审查 + references.bib |
process-journal 当前已归档到 deprecated/skills/,等待 journal acquisition 重新设计。
| Agent | 职责 |
|---|---|
search-agent | 将研究意图转成 `quasi-search book |
download-agent | 文件获取、候选判断、接受入库 |
extract-agent | EPUB/PDF/OCR/章节切分编排 |
analyse-agent | 论文/章节/讲座转写分析 |
synthesis-agent | book/author/topic 综合报告 |
audit-agent | vault consistency 检查和可修复项处理 |
proofread-agent | draft 局部校对 |
citecheck-agent | 引文 context-fit 审查 |
translate-agent | 双语翻译 |
quasi-search book|paper ...
quasi-download book candidates|fetch ...
quasi-download paper fetch ...
quasi-download accept ...
quasi-extract epub|ocr|split ...
quasi-transcribe run|classify|silent ...
quasi-audit --path ...
quasi-helpers proofread prepare|cleanup ...
quasi-helpers citation parse|biblio|resolve|review-cards|emit-bib ...
quasi-helpers localise scan|write ...
quasi-translate ...
旧 quasi-citation 和 quasi-proofread bin 已移除;新流程走
quasi-helpers 和 quasi-audit。
vault/
books/{slug}/
papers/{slug}.md
authors/{slug}.md
drafts/
sources/
{slug}.{epub,pdf}
.quasi/
audit/
citation/
localise/
proofread/
temp/
通过 /plugin → Configure 填入:
| 服务 | 配置字段 |
|---|---|
| Anna's Archive | anna_donator_key |
| CookieCloud / EZProxy | cookiecloud_server, cookiecloud_uuid, cookiecloud_password, cookiecloud_ezproxy_domain, cookiecloud_ezproxy_base_url |
| Immersive Translate | immersive_auth_key |
| Kagi | kagi_session_token |
| Google Scholar proxy | google_scholar_proxy_url |
npx claudepluginhub giraphant/quasi --plugin quasiA research infrastructure for AI agents. Search, read, and analyze papers from your local knowledge base while coding. Includes arXiv discovery, layered reading, ingestion, topic modeling, citation graphs, insights analytics, Office document inspection, scientific tool docs, and academic writing workflows. Requires Python 3.10+ and pip install.
Simulate peer review by constructing reviewer personas from Zotero sources. Identifies relevant perspectives, retrieves full texts, builds reviewer profiles, and generates focused reviews on theory/methods and findings.
深度研究 — 13 agent 协作:研究问题定义、系统性检索、偏差评估、综合分析、引用编译
Academic literature search, citation management, PDF retrieval, and literature review synthesis
Collection of academic skills based on effortlessacademic.com note taking ideas like atomic sentences.
Use when the user needs academic paper retrieval — searching scientific literature by author/year/journal, finding paper chunks for RAG-style citations, or expanding original text around a known paper offset. Provides three Sciverse tools (search_papers, semantic_search, read_content) via the sciverse-mcp-server MCP server.