Skillget

discolike-skills

mcp
Guvenlik Denetimi
Uyari
Health Uyari
  • License — License: MIT
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Low visibility — Only 5 GitHub stars
Code Gecti
  • Code scan — Scanned 9 files during light audit, no dangerous patterns found
Permissions Gecti
  • Permissions — No dangerous permissions requested

Bu listing icin henuz AI raporu yok.

SUMMARY

Skills that teach coding agents to find companies, enrich them, and reach contacts with DiscoLike. Claude Code plugin marketplace included.

README.md

DiscoLike

Official DiscoLike skills for coding agents
The GTM search engine for the business web, one workflow for Claude Code, Codex, Cursor, Gemini CLI, Copilot, Windsurf, Grok, and agents that honor AGENTS.md.

Skill: discolike Plugin version License: MIT

Website · MCP server · API Docs · Python SDK & CLI · Sign up · Book a demo

DiscoLike is a search engine over 80M+ business websites worldwide, analyzed in 55 languages; roughly 3x the coverage of LinkedIn-dependent databases. It ranks companies by what they actually do on the web, so it reaches niche verticals, local businesses, and non-English markets that other B2B data sources miss.

This repo ships the discolike plugin: the discolike skill with its flow and troubleshooting references, and setup and feedback skills. The plugin ships no executables; shell commands the skills suggest go through the client's normal permission prompt. The main skill gives a coding agent the operating knowledge to use DiscoLike well: which access mode to pick, how to open an account without a browser, how to search without wasting credits, and the end-to-end flows for the jobs people actually ask for; target account lists, lookalikes, CRM enrichment, ICP segmentation, and contact discovery.

If you're a coding agent: read SKILL.md directly; it is the whole product. The fastest integration is the hosted MCP server at https://api.discolike.com/v1/mcp (streamable-http, OAuth). For scripting use pip install discolike; for the terminal use pip install discolike-cli or uvx --from discolike-cli discolike. No account yet? POST https://api.discolike.com/v1/public/signup opens one from an email and name; the owner confirms by email. Machine-readable API index: https://docs.discolike.com/llms.txt.

Installation

Claude Code

Add the marketplace once, then install the plugin:

/plugin marketplace add Discolike/discolike-skills
/plugin install discolike@discolike

The skill activates automatically when a task involves finding companies, building a target or TAM list, lookalikes, ICP research, company enrichment, or decision-maker contacts.

Any agent with the skills CLI

Works for Codex, Cursor, Gemini CLI, Copilot, Windsurf, OpenCode, and Claude Code alike:

npx skills add Discolike/discolike-skills

Codex, Cursor, and other agents

Clone this repository into the project or install it as a plugin where the client supports plugins. Every adapter points back to the same canonical skills/discolike/SKILL.md, so clients get the same safeguards and product knowledge without maintaining forks.

Client Included integration How it activates
Claude Code .claude-plugin/ plus .claude/CLAUDE.md Install from the Claude marketplace, or open the repo as a project.
Codex .codex-plugin/plugin.json plus AGENTS.md Install the plugin from a marketplace that contains this repo, or open the repo; AGENTS.md routes matching work to the skill.
Cursor .cursor-plugin/plugin.json plus .cursor/rules/discolike.mdc Install from the Cursor Marketplace, or open or copy the repository rule into a Cursor project.
Gemini CLI gemini-extension.json plus GEMINI.md gemini extensions install https://github.com/Discolike/discolike-skills, or open the repo or copy the file into a project.
GitHub Copilot .github/copilot-instructions.md Open or copy the instruction file into a project.
Windsurf .windsurf/rules/discolike.md Open or copy the repository rule into a Windsurf project.
Grok .grok-plugin/plugin.json plus AGENTS.md Install the plugin where supported, or use the included project instruction file.
Other agents AGENTS.md Use the included project instruction file, or add its two paragraphs to the agent's project instructions.

For a personal global Codex installation, fetch the canonical skill directly:

mkdir -p ~/.agents/skills/discolike
curl -fsSL https://raw.githubusercontent.com/Discolike/discolike-skills/main/skills/discolike/SKILL.md \
  -o ~/.agents/skills/discolike/SKILL.md

If a client does not support rules or skills, give it the contents of AGENTS.md as project instructions. The agent will then load the canonical skill only for relevant tasks instead of carrying the full workflow in every prompt.

What the plugin bundles

Piece Purpose
skills/discolike/ Entry-point skill plus flows.md and troubleshooting.md, loaded on demand
skills/setup/, skills/feedback/ Connect and report

What the skill knows

  • Which door to use. MCP server for clients that support remote MCP, CLI when shelling out, Python SDK when writing a script, REST as a last resort; and when to ask the user before wiring any of them up.
  • How to open an account from the agent. One POST with an email and name, no browser, no credential returned; the owner confirms by email and picks a plan.
  • How to search. The three ways to describe a target on one discover call; a plain-English icp_prompt, up to 10 seed domains for lookalikes, exact phrase_match fragments; plus the geo, size, revenue, tech-stack, category, language, and business-model filters, each with a negate_ twin.
  • How not to burn credits. Count first (free), start with a small max_records, and page past the 10,000-per-call cap with exclusion lists rather than re-running the same search.
  • What to do after discovery. Contacts at matched companies, enrichment of lists you already have, segmenting a customer list into ICP clusters, validating a list against an ICP, and pushing results to HubSpot, Salesforce, or Pipedrive.
  • How to diagnose bad results. What to change when results look noisy, and when a question is a research job rather than a filter.

Flows

Contacts and prospecting flows use MCP. Other flows include CLI and SDK examples where available.

Flow What the agent does
Discover, verify, contacts, emails Build a target list, confirm company fit, then deliver Validated and Validated (catch-all resolved) contacts with validated work emails
Enrich a CRM export Append firmographics, tech stack, growth, and scores to a domain list
Match company names to domains Turn a messy name column into resolved domains, single or bulk
Segment a client list into ICPs Cluster existing customers, describe each cluster, run lookalikes per cluster
Market map to N Grow a list to a target size with exclusion-list paging
ICP from a website Derive a target profile from one URL and search on it
Signal-qualified list "Find X that run Shopify", "that are hiring SDRs", "that have a pricing page"; discover, then qualify on the signal
Technology and infrastructure targeting Find companies by the vendors and stack they run
Rank or filter a list the user already has Score an existing list against an ICP instead of discovering new companies

Data and privacy

The plugin talks to these services, and only these:

Service When What is sent
api.discolike.com Every MCP, CLI, SDK, or REST call, and account signup Search filters, domains, uploaded lists, and the email and name given at signup
GitHub (github.com) feedback skill: the user opens a prefilled issue link and submits it themselves The issue title and body the user approved, as a public issue in a Discolike repository
discolike.com, docs.discolike.com When the agent reads DiscoLike documentation Page requests; no user data

Results that include contacts (names, emails) are written to local files the agent creates, and saved queries and exclusion lists are stored in the user's DiscoLike account. See the privacy policy for retention.

Related

Support & contact

License

MIT

Yorumlar (0)

Sonuc bulunamadi