Atlas / MCP servers / dralkh / Seerai

SeeraiBLOCK

Name: Seerai
Author: dralkh

mcp/dralkh/seerai

MCP server by dralkh · academic research · agent · ai · Audited 2026-10-07

Zotero AI plugin Research assistant for Zotero 10. Chat with your library, run federated scholarly search, RAG, OCR, systematic reviews, and manage cloud storage. Includes standalone MCP, Agentic capabilities, and skills library.

Verdict

BLOCK

Grade

Trust score

20 /100

Exposed tools

54 36r · 16w · 2d

Transport

stdio

License

MIT

Stars

Overview

From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.

seerai is an intelligent research assistant plugin for Zotero 9/10 that integrates AI-powered chat, advanced search, structured data extraction, and systematic review workflows directly into your research workflow. Chat with your papers, extract structured data, run PRISMA-style reviews, and accelerate your literature review with a local-first, privacy-focused design.

Installation

From GitHub (Recommended)

Download the latest release (.xpi file) from Releases.
In Zotero, go to Tools → Add-ons.
Click the gear icon ⚙️ and

Read from source at commit d338c4197636OBSERVED · 2026-10-07

Connect

Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.

claude-code

claude mcp add seerai --env ANTHROPIC_API_KEY=${ANTHROPIC_API_KEY} --env EXA_API_KEY=${EXA_API_KEY} --env FOUNDRY_API_KEY=${FOUNDRY_API_KEY} --env MP_API_KEY=${MP_API_KEY} -- npx -y [email protected]

claude-desktop

{
  "mcpServers": {
    "seerai": {
      "command": "npx",
      "args": [
        "-y",
        "[email protected]"
      ],
      "env": {
        "ANTHROPIC_API_KEY": "${ANTHROPIC_API_KEY}",
        "EXA_API_KEY": "${EXA_API_KEY}",
        "FOUNDRY_API_KEY": "${FOUNDRY_API_KEY}",
        "MP_API_KEY": "${MP_API_KEY}"
      }
    }
  }
}

Exposed tools (54)

36 read · 16 write · 2 destructive. Blast radius: 2 tools can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.

Tool	Risk	Description
`check_environment`	read	Check available runtimes and tools on the host computer. Reports Python, Node, Git, pip/npm versions, and shell availability. Use before installing packages or running skill scripts.
`citation-management`	write	Validate citations, organize references, manage Zotero collections, and create citation-backed notes.
`clarify`	read	PAgent-compatible clarification tool.
`collection`	destructive	Manage Zotero collections. Find, create, list contents, or add/remove items from collections.
`context`	write	Manage conversation context. Add items (papers, tags, collections) to focus the conversation, remove them, or list current context.
`delegate_task`	write	Run a bounded non-tool sub-agent task through the plugin.
`document-processing`	read	Process PDFs, DOCX, markdown, images, and workspace files using SeerAI OCR, document conversion, and file tools.
`execute_code`	write	Run a temporary Python, JavaScript, or Bash snippet.
`experimental-design-and-power`	read	Support study design, protocol criteria, outcome selection, effect measures, and power-analysis planning.
`generate_item_tags`	read	Generate AI-powered tags for a Zotero item based on its content.
`generate_tags`	read	Generate structured tags for an academic paper
`get_item_metadata`	read	Get complete metadata for a Zotero item (authors, DOI, abstract, etc.).
`import_paper`	write	Import a paper from a federated scholarly corpus into Zotero. Accepts search_external paper IDs and common arXiv, DOI, PubMed, repository, preprint, and Semantic Scholar identifiers.
`keyword_search`	read	Fast keyword-based search of your Zotero library using BM25 lexical matching.
`literature-review`	write	Plan, screen, synthesize, and write literature reviews with Zotero collections, RAG, tables, and systematic review workflows.
`mixture_of_agents`	write	Run bounded non-tool sub-agent perspectives and synthesize.
`note`	write	Create or edit Zotero notes. Create new notes attached to items/collections, or edit existing notes with operations.
`paper-lookup`	write	Find, inspect, import, and read academic papers using Zotero, Semantic Scholar, web search, related-paper discovery, and OCR-backed content reading.
`patch`	write	Apply a targeted fuzzy patch to a workspace file.
`peer-review`	read	Review papers or manuscripts for novelty, validity, reporting quality, methods, limitations, and actionable revisions.
`process`	read	Manage background terminal processes.
`read_chunks`	read	Read specific text chunks/passages from indexed papers by chunk ID or item ID.
`read_file`	read	Read a workspace text file with pagination.
`read_item_content`	read	Read the full content of an existing Zotero paper including notes and PDF text. Numeric item_id values are Zotero item IDs; string values are read-only aliases for existing Zotero items.
`related_papers`	read	Find related papers via citations (forward) or references (backward).
`scientific-writing`	read	Draft, revise, and structure scientific prose grounded in Zotero evidence and verified citations.
`search_external`	read	Search external scholarly corpora. Defaults to Semantic Scholar; supports smart modes and specific corpora such as PubMed, arXiv, Europe PMC, CORE, BASE, Zenodo, HAL, bioRxiv, medRxiv, and IACR.
`search_files`	read	Search workspace filenames and file contents.
`search_library`	read	Search the Zotero library for papers matching a query. Returns titles, authors, and IDs.
`search_similar`	read	Find papers similar to a given item using embedding similarity.
`semantic_search`	read	Semantically search your Zotero library for relevant passages.
`skill_info`	read	Get filesystem information about a bundled skill: absolute skill directory path, available scripts, references, and assets. Use before executing skill scripts so you know the correct paths.
`skill_manage`	read	Manage local Agent Skills registry state and trusted sources.
`skill_reference`	read	Read a reference or script file from a bundled skill directory. Use this to access API documentation, configuration templates, or utility scripts included with a skill.
`skill_view`	read	View and activate the full instructions for one Agent Skill.
`skills_list`	read	List and search SeerAI Agent Skills.
`statistical-analysis`	read	Guide extraction and interpretation of effect sizes, confidence intervals, heterogeneity, and evidence synthesis.
`systematic_review`	write	Read and update systematic review projects, analyze papers, save grounded extraction data, and run or inspect evidence synthesis and gap analysis.
`table`	write	Manage research analysis tables. List tables, create new tables, add papers, add columns, generate AI data, complete PDF-aware generation, or read table contents. Prefer action=
`task_complete`	read	Signal that the current task is fully complete. Call this ONLY when ALL requested work has been DONE.
`terminal`	write	Execute a shell command in the workspace through guarded MCP execution.
`todo`	read	PAgent-compatible TODO wrapper over todoread/todowrite.
`todoread`	read	Read the current TODO list for this session. Call this when you lose track of what remains to be done or after context compaction to recover your task state.
`todowrite`	write	Create or update a structured task list (TODO list) for the current session.
`web`	read	Search the web or read webpage content.
`workspace_diff`	read	Show the diff (changes) for a workspace file between versions.
`workspace_edit_file`	write	Edit a file in the workspace by replacing an exact string match with new content.
`workspace_glob`	read	Find files in the workspace matching a glob pattern (e.g.
`workspace_grep`	read	Search for a regex pattern across files in the workspace.
`workspace_log`	read	Show version history for a workspace file.
`workspace_question`	read	Ask the user one or more questions during execution. Returns the user
`workspace_read_file`	read	Read a file from the workspace. Returns the file content and metadata.
`workspace_write_file`	write	Write a file to the workspace. Creates the file if it does not exist, overwrites if it does.
`write_file`	destructive	Write or overwrite a workspace file completely.

Trust audit

BLOCKgrade F · trust 20/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.

Layer	What it checks	Result
L0	Provenance & inventory	PASS
L1	Static analysis of the code	WARN
L2	Instruction surface (what it tells the agent)	FAIL
L3	Class-specific surface	WARN
L4	Behavioural (sandbox)	SKIPPED

What the source does

Filesystem: declared (7 observation(s))
Network: declared (9 observation(s))
Shell: declared (3 observation(s))
Dependencies: not all pinned
Secrets in source: found

Findings (25)

CRITICALPrompt injection · prompt.transfer_instruction · CWE-94, CWE-1427

skills/labarchive-integration/SKILL.md:199

python3 scripts/notebook_operations.py backup-all --email [email protected] --password AUTH_TOKEN

Why it matters. an instruction to move sensitive data to an outside destination

Fix. remove; a skill never needs the user's secrets off the machine

HIGHPrompt injection · prompt.credential_read · CWE-94, CWE-1427

skills/adaptyv/SKILL.md:22

When writing code, always read the API key from the environment variable `ADAPTYV_API_KEY` or from a `.env` file — never hardcode tokens. Check for a `.env` file in the project root first; if one exis

Why it matters. asks the agent to read credentials

HIGHPrompt injection · prompt.credential_read · CWE-94, CWE-1427

skills/autoskill/SKILL.md:14

> **Network access & environment variables.** This skill makes authenticated HTTP requests to (a) the user's local screenpipe daemon on loopback, and (b) the user-configured LLM backend — one of `http

Why it matters. asks the agent to read credentials

HIGHPrompt injection · prompt.credential_read · CWE-94, CWE-1427

skills/benchling-integration/SKILL.md:434

- Read only named environment variables (`BENCHLING_TENANT_URL`, `BENCHLING_API_KEY`, etc.)

Why it matters. asks the agent to read credentials

HIGHPrompt injection · prompt.credential_read · CWE-94, CWE-1427

skills/biopython/SKILL.md:54

For NCBI database access, always set your email address (required by NCBI). For reusable software, set a stable `Entrez.tool` value and register the tool/email with NCBI. For higher rate limits (10 re

Why it matters. asks the agent to read credentials

MEDIUMCode injection · code.dynamic_import · CWE-78, CWE-94, CWE-95

skills/diffdock/scripts/setup_check.py:40

module = __import__(import_name)

MEDIUMCode injection · code.dynamic_import · CWE-78, CWE-94, CWE-95

skills/timesfm-forecasting/scripts/check_system.py:362

mod = importlib.import_module(import_name)

MEDIUMNetwork egress · net.beacon_words · CWE-200, CWE-319

src/modules/systematicReview/utils.ts:222

"beacon",

MEDIUMNetwork egress · net.raw_ip · CWE-200, CWE-319

addon/prefs.js:148

pref("execServerUrl", "http://127.0.0.1:23120");

MEDIUMNetwork egress · net.raw_ip · CWE-200, CWE-319

mcp-server/src/index.ts:232

`[seerai] HTTP execution API listening on http://127.0.0.1:${execPort}`,

MEDIUMNetwork egress · net.raw_ip · CWE-200, CWE-319

mcp-server/src/index.ts:245

res.setHeader("Access-Control-Allow-Origin", "http://127.0.0.1:23119");

MEDIUMObfuscation / stealth · obf.base64_blob · CWE-506, CWE-94

skills/diffdock/assets/batch_template.csv:3

example_2,,COc1ccc(C#N)cc1,MSKGEELFTGVVPILVELDGDVNGHKFSVSGEGEGDATYGKLTLKFICTTGKLPVPWPTLVTTFSYGVQCFSRYPDHMKQHDFFKSAMPEGYVQERTIFFKDDGNYKTRAEVKFEGDTLVNRIELKGIDFKEDGNILGHKLEYNYNSHNVYIMADKQKNGIKVNFKIRHNIED

MEDIUMObfuscation / stealth · obf.base64_blob · CWE-506, CWE-94

skills/diffdock/scripts/prepare_batch_csv.py:172

'MSKGEELFTGVVPILVELDGDVNGHKFSVSGEGEGDATYGKLTLKFICTTGKLPVPWPTLVTTFSYGVQCFSRYPDHMKQHDFFKSAMPEGYVQERTIFFKDDGNYKTRAEVKFEGDTLVNRIELKGIDFKEDGNILGHKLEYNYNSHNVYIMADKQKNGIKVNFKIRHNIEDGSVQLADHYQQNTPIGDGPVLLPDNH

MEDIUMObfuscation / stealth · obf.base64_blob · CWE-506, CWE-94

skills/glycoengineering/SKILL.md:88

fc_sequence = "APELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPEVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISKAKGQPREPQVYTLPPSREEMTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSKLTVDKS

MEDIUMObfuscation / stealth · obf.homoglyph · CWE-506, CWE-94

skills/neuropixels-analysis/scripts/neuropixels_pipeline.py:133

ax.set_ylabel('Depth (μm)')

MEDIUMObfuscation / stealth · obf.homoglyph · CWE-506, CWE-94

skills/neuropixels-analysis/scripts/neuropixels_pipeline.py:143

print(f"  Estimated drift range: {drift_estimate:.1f} μm")

MEDIUMObfuscation / stealth · obf.homoglyph · CWE-506, CWE-94

skills/neuropixels-analysis/scripts/neuropixels_pipeline.py:383

print(f"Drift > 20 μm detected, applying motion correction...")