The measured layer for AI capabilities — search, rank, and install MCP servers & agent skills from your terminal. Reads live public data from tashan.sh. No account, no backend, no telemetry.

npx tashan-cli search database        # find capabilities, ranked by Trust
npx tashan-cli top browser            # the leaderboard, optionally by category
npx tashan-cli info context7          # the measured dossier for one capability
npx tashan-cli add context7           # the install command for your client   ← the money shot

Commands

command	does
`tashan search <query>`	ranked matches by name / package / category
`tashan top [category]`	top capabilities by Trust
`tashan info <name>`	Trust, maintenance, vitality, expertise verdict, adoption, links
`tashan add <name>`	the exact install command — same as the website

Flags

--client <c> — claude · cursor · desktop · codex · npx (for add)
--json — machine-readable output (pipe into jq)
--limit <n> — cap the rows (default 20)

tashan add supabase --client cursor        # print the mcp.json block for Cursor
tashan top --json | jq '.[0].trust'        # scriptable

How it works

One file, zero dependencies, Node ≥ 18. Every score is re-derivable from public evidence; the ranking can't be bought. The install commands are generated by the same logic as the website, so the terminal and the page never disagree.

TASHAN_DATA=<path> reads a local index.json instead of the network (dev/offline); TASHAN_SITE=<url> points at a different deployment.

node tashan.test.mjs        # run the pure-logic tests (no network)

Keywords

mcp

model-context-protocol

claude

cursor

agent-skills

FAQs

What is tashan-cli?

Is tashan-cli well maintained?

Package last updated on 08 Aug 2026

Did you know?

Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.

Install