Atlas / MCP servers / jordan-gibbs / hyperresearch

hyperresearchCAUTION

Name: hyperresearch
Author: jordan-gibbs

mcp/jordan-gibbs/hyperresearch

MCP server by jordan-gibbs · agents · agentskills · citations · Audited 2026-10-10

Convert Claude Code or Codex into the most intelligent Deep Research Agent. Collect, search, and synthesize web research into a persistent, searchable wiki that builds on itself. Hosted API + MCP: hyperresearch.ai

Verdict

CAUTION

Grade

Trust score

87 /100

Exposed tools

13 9r · 4w · 0d

Transport

streamable-http

License

MIT

Stars

3,824

Overview

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

The Most Powerful Deep Research Harness: a deep research skill for Claude Code and OpenAI Codex

Hyperresearch turns Claude Code into a deep research agent, and it's #1 on the DRACO deep research benchmark, ahead of Claude Research, ChatGPT, Gemini and Perplexity. It runs in OpenAI Codex too. A tier-adaptive 16-step pipeline takes one prompt and produces an adversarially-audited report with full source provenance. Every source it reads lands in a persistent, searchable vault, so each session starts smarter than the last.

Read from source at commit c793941599fbOBSERVED · 2026-10-10

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 hyperresearch --env EXA_API_KEY=${EXA_API_KEY} --env OPENAI_API_KEY=${OPENAI_API_KEY} --env SERPLY_API_KEY=${SERPLY_API_KEY} --env TAVILY_API_KEY=${TAVILY_API_KEY} -- uvx hyperresearch

claude-desktop

{
  "mcpServers": {
    "hyperresearch": {
      "command": "uvx",
      "args": [
        "hyperresearch"
      ],
      "env": {
        "EXA_API_KEY": "${EXA_API_KEY}",
        "OPENAI_API_KEY": "${OPENAI_API_KEY}",
        "SERPLY_API_KEY": "${SERPLY_API_KEY}",
        "TAVILY_API_KEY": "${TAVILY_API_KEY}"
      }
    }
  }
}

Exposed tools (13)

9 read · 4 write · 0 destructive.

Tool	Risk	Description
`check_source`	read	Check if a URL has already been fetched into the research base.
`create_note`	write	Create a new research note.
`fetch_url`	write	Fetch a URL and save it as a research note.
`get_backlinks`	read	Get all notes that link TO a given note.
`get_hubs`	read	Get the most-linked-to notes in the research base (hub notes).
`lint_vault`	write	Run health checks on the vault. Returns issues found.
`list_notes`	read	List notes with optional filters. Returns summaries (no bodies).
`list_sources`	read	List fetched web sources, optionally filtered by domain.
`read_many`	read	Read multiple notes at once. Pass comma-separated IDs.
`read_note`	read	Read a single note by ID. Returns full metadata and body content.
`search_notes`	read	Search the research base by text. Returns matching notes with titles, summaries, and full bodies.
`update_note`	write	Update a note
`vault_status`	read	Get vault health overview: note counts, tag distribution, link stats, word count.

Trust audit

CAUTIONgrade B · trust 87/100 Install with care. The audit found things worth knowing before you trust its output.

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

What the source does

Filesystem: declared (4 observation(s))
Network: declared (11 observation(s))
Shell: declared (1 observation(s))
Dependencies: pinned
Secrets in source: none-found

Findings (19)

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

src/hyperresearch/scholar/registry.py:40

module = importlib.import_module(f"hyperresearch.scholar.providers.{module_suffix}")

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

src/hyperresearch/serve/server.py:617

url = f"http://127.0.0.1:{port}"

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

tests/test_core/test_claims_and_embed.py:274

escaping_tag = os.path.relpath(outside, runs)  # ../../..\\outside-vault

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

tests/test_core/test_note.py:263

path = write_note(tmp_vault.notes_dir, "Evil", note_id="../../../outside")

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

tests/test_core/test_patterns.py:279

assert not is_valid_wiki_link_target("../../../tmp/pwned")

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

tests/test_core/test_patterns.py:280

assert not is_valid_wiki_link_target("a/../../b")

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

tests/test_core/test_run_tag_validation.py:41

"../../x",

LOWNetwork egress · net.metadata_ip · CWE-200, CWE-319

tests/test_core/test_oa_core_resolver.py:470

rec = _record(id=None, downloadUrl="http://169.254.169.254/latest/meta-data", sourceFulltextUrls=[])