taxueseek/argo
Search built for agents: multilingual coverage across web, academic, code, shopping, finance, news, and encyclopedias.
Install
via npm
Resolves a published tarball rather than cloning the repository, and installs without any extra setup. Swap `web` for your profile name if you run another one.
via GitHub · npm package
Installing from GitHub runs the project's build script, which pnpm blocks until you allowlist it — run the command once and pnpm prints the exact key to add under `allowBuilds` in ~/.dsh/profiles/web/pnpm-workspace.yaml.
What happened when we ran it
Installs, but not as a profile layer
The command works. The package declares no dsh.bundle, so the harness takes it as an ordinary dependency rather than a layer over your profile — use it the way its README describes. If a later version adds one, it activates on its own.
Every command here is run in a throwaway container against a clean profile, and the result is whatever the harness recorded — not a guess from the source. Last run yesterday.
Due diligence
What it is — Argo is a multilingual search infrastructure for AI agents, delivering structured evidence candidates with credibility decomposition.
Who it is for — If you are developing AI agents using DeepSeek Harness to query topics like finance, sports, or movies requiring structured evidence support, it fits. If you seek only human-readable search summaries or link lists, you do not need it.
Watch out — Successfully installed via command, but treated as ordinary dependency without dsh.bundle so no profile layer added. No obvious issues found in sandbox testing.
The verdict — I would install it because it offers structured evidence for agents to reduce context bloat and costs, but as an ordinary dependency it will not add a profile layer and must be used as per its documentation.
Generated by grok-4.6, and a starting point rather than a verdict. Where it says a plugin installs or does not, that is from a real run in a clean profile — everything else is read off the repository. Trust the source over this.
argo is a DeepSeek Harness plugin that provides multilingual search and evidence-oriented results for web, academic, code, shopping, finance, news and encyclopaedia queries. It is intended for agents that need compact, sortable material rather than a long generated answer or an unstructured list of links.
The plugin exposes Argo through MCP, providing 10 mcp__argo__* tools. A query can pass through language detection, domain routing, multiple search engines, RRF result fusion and a fast evidence review. The returned JSON includes fields such as selection, absorption, freshness, credibility_fast, evidence_flags, engine_outcomes and recovery. Argo also supports search.py for command-line use, mcp_server.py as a local Python MCP entry point, and local file search. It uses an in-memory and SQLite cache; configuration is read with PyYAML.
This plugin is for research, fact checking and domain-specific retrieval where language and source choice matter. It is a poor fit when only a simple search command is needed, or when browser rendering, screenshots or PDF extraction are required without their optional dependencies. Python 3.10+ is required, and API keys are optional rather than guaranteed for every source. With the listed terminal-surface risk, the plugin reaches the agent’s terminal to run its Python scripts, MCP server and optional local search tools; that access supports command-line and local-file search, but should be considered when granting it to an agent.
Argo routes each query using language detection, domain rules, TF-IDF, budget and cached hot paths. It can query several engines in parallel, combine results with RRF, optionally rerank them, and perform a fast evidence review. Empty results can trigger staged recovery: widening the query, switching to a related or general engine, and then trying another language. Results are returned as compact JSON, with engine_outcomes and recovery included when applicable.
The evidence fields are intended for agent-side sorting. selection reflects source authority, absorption reflects evidence density, and freshness reflects publication timing. The README recommends search followed by fetch for high-consequence questions, and advises treating search-result pages, redirect chains and social posts cautiously.
| Entry point | Purpose |
|---|---|
scripts/search.py "query" --json |
Run a JSON search from the command line |
scripts/search.py --list-engines |
List available engines |
scripts/mcp_server.py |
Start the local Python MCP server |
npx -y github:taxueseek/argo |
Start the GitHub MCP entry point |
from search import super_search |
Call the library from Python |
The DSH bundle is under packages/dsh-plugin/. A user-level cordis.patch.yml can override the mcp-argo source.
PyYAML reads the configuration files. API keys are optional: without them, Argo uses free engines and local local_* engines, while configured sources can provide additional coverage. The optional environment variable ARGO_PYTHON selects the Python executable for the npx entry point.
Python 3.10+ and PyYAML are required. Node.js 18+ is needed for the npx MCP route. curl_cffi, ddgs CLI, realtime-index CLI, Chrome, pdfplumber, PyMuPDF and Playwright are optional. They add browser-TLS handling, local search backends, real-time indexing, rendering or screenshots, and PDF extraction.
Without optional dependencies, the related engine or tool is disabled or falls back as documented. In particular, argo_pdf requires a PDF extractor, and screenshot support requires Chrome or Playwright. The README does not name individual API-key configuration keys.
Written from the project's own documentation and kept in sync with it. Where the two disagree, the source is authoritative — read the README on GitHub
Same category
superdesigndev
Tool catalog for agents: search ~2,600 external endpoints (SEO and SERP, backlinks, social, people and company enrichment, ad libraries, scraping) by the task you want done, read each one's parameters and per-call price, then call it with the credential injected server-side. Ships the skill plus an MCP row that stays disabled until TREG_TOKEN is set.
Installed cleanly when we ran it
Lum1104
Chrome sidebar extension that lets DSH operate your browser directly, no vision capabilities required.
Installed cleanly when we ran it
zhaoolee
Export DSH conversations as Smartisan Notes-style PNGs, or create and update Markdown notes in a configured account-scoped workspace.
Installed cleanly when we ran it
liustack
Web search bridge for text-only agents: ask the web or X, get structured JSON evidence (search, fetch, citations).
Installed cleanly when we ran it
lire1131
Undo/redo & rollback system for DSH: every config change is auto-snapshotted; undo/redo/restore to any version from the WebUI or the offline CLI/GUI tools (works even when DSH fails to boot).
Installed cleanly when we ran it
Vladimir-Human
Skills and optional MCP rows for ten Russian marketplaces: price comparison across Wildberries, Detsky Mir and Yandex Market, plus per-source search, product cards and reviews. The 13 skills load on install; both MCP rows stay disabled until RU_MARKETPLACE_MCP_DIR points at a local clone, which needs Python 3.12+ and uv.
No one-line install — this plugin lives inside a larger repository and publishes no npm package.
