Is a Claude plugin the same as an MCP server?

No. MCP connects Claude to live external data. A plugin or skill bundle teaches Claude a repeatable workflow and guardrails for using that data.

Should a plugin contain Peec credentials?

No. Credentials belong in OAuth, approved secrets storage, or the MCP client's secure configuration, not in plugin files.

What should the Peec plugin do first?

Resolve the project, brands, topics, tags, and model channels so the report uses human-readable names.

Can a plugin trigger Peec write tools?

It can describe approved workflows, but any write or delete operation should require explicit human confirmation.

What is the best first plugin workflow?

Start with a read-only weekly visibility pulse or source authority audit before adding configuration-management workflows.

Peec AI Claude Plugin MCP Claude Agents GEO

How to Build a Claude Plugin with Peec AI MCP

Marco Lobo

28 May 2026·Updated 28 May 2026·7 min read

Last updated: 28 May 2026

Short answer: a Claude plugin should not replace the Peec AI MCP server. The plugin should package the operating instructions, workflows, guardrails, examples, and QA checks. Peec AI MCP remains the live data connector.

Peec AI MCP blog visual from Peec's first-party launch article

Think of the plugin as the agent's playbook. It tells Claude how to use Peec data: default date ranges, metric formatting, action categories, confirmation rules, output templates, and what counts as a finished report. MCP supplies the actual project data.

Plugin anatomy

File or section	Purpose
Manifest	Names the plugin and declares what it does
Skill instructions	Teach Claude the Peec reporting workflow
Examples	Show weekly pulse, source audit, and blog optimisation prompts
Guardrails	No secrets, no raw IDs, no writes without approval
QA checklist	Schema, links, images, source ledger, desktop/mobile preview

What the plugin should teach Claude

Start by listing Peec projects and resolving the right one by name.
Resolve brands, topics, tags, and model channels before analysis.
Use 30 days for strategy, 7 days for weekly reporting.
Break down by engine and topic, not just aggregate.
Display visibility and share of voice as percentages.
Display retrieval rate and citation rate as rates.
Group actions into owned, editorial, reference, and UGC work.
Confirm before using any write or delete tool.

Example plugin workflow

Run a Peec source authority audit for <project>.
Find gaps where competitors are cited and our domain is not.
Return: executive summary, top engines, top topics, source domains, URLs,
recommended owned-page updates, editorial targets, UGC opportunities,
and the exact human approvals needed before any mutation.

Plugin versus connector

The connector/MCP server answers "what data can Claude access?" The plugin answers "how should Claude behave with that data?" Mixing those boundaries is risky. Do not hardcode Peec tokens in a plugin. Do not assume every workspace has the same project names. Do not hide destructive actions behind vague wording.

For Schmitdy, this plugin pattern would sit beside Claude Code implementation, Cowork reporting, and the AI Search audit.

Source and fact ledger

Peec AI MCP tools reference: https://docs.peec.ai/mcp/tools
Peec AI MCP setup guide: https://docs.peec.ai/mcp/setup
Claude connectors directory docs: https://claude.com/docs/connectors/directory

Frequently Asked Questions

Peec AI Claude Plugin MCP Claude Agents GEO

Marco Lobo

Founder, Schmitdy

Marco builds AI search growth systems that turn prompts, sources, content, and agents into pipeline for B2B teams.

LinkedIn X / Twitter

First-party Peec AI MCP launch visual explaining how AI visibility data connects to agent workflows

AI GuidesPeec AIMCP

Peec AI MCP Deep Dive: How AI Search Data Becomes Agentic GEO Execution

Peec AI MCP turns AI visibility data into an agent-readable data layer. The winning workflow is not dashboard replacement; it is measured, source-led execution.