mcp-plesk-dev-docs copilot-instructions.md

Project instructions for a unified Plesk documentation search server. Plesk is a hosting-control product, and the server finds relevant pages and code references from its documentation.

In plain words
What is it for?
Use it when developing or maintaining the MCP server, its document parsers, vector index, embeddings, reranking, source synchronization, or retrieval behavior.
Why use it?
It explains how documentation is collected, divided into searchable sections, indexed, ranked, and returned so contributors can work within the existing design.

Instructions file for GitHub Copilot

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/barateza/mcp-plesk-dev-docs/copilot-instructions
Clone the repo
git clone --depth 1 https://github.com/barateza/mcp-plesk-dev-docs

Made for: GitHub Copilot.

Per session 2,967 This file is loaded in full into every session.
When invoked 2,967 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.02967 $0.02967
Opus 5 $0.01484 $0.01484
Sonnet 5 $0.00593 $0.00593
Haiku 4.5 $0.00297 $0.00297

Measured 2d ago against content hash cf9788c0a73a, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

mcp-plesk-dev-docs copilot-instructions.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.github/copilot-instructions.md · 281 lines

How it starts

The opening of the file, as written. The whole thing — 281 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Plesk Unified MCP Server - AI Agent Instructions

Project Overview

Plesk Unified is Model Context Protocol (MCP) server that indexes and retrieves Plesk docs via semantic search w/ vector embeddings and intelligent reranking. It aggregates docs from multiple sources (API, CLI, Admin Guide, PHP stubs, JS SDK) into unified knowledge base powered by LanceDB.

Architecture & Components

Core Stack

  • Server Framework: FastMCP (Model Context Protocol)
  • Embeddings: BAAI/bge-m3 (multilingual, bidirectional)
  • Reranking: BAAI/bge-reranker-base (cross-encoder for result ranking)
  • Vector DB: LanceDB w/ Apache Arrow backend
  • Parsing: BeautifulSoup4 for HTML, raw text for PHP/JS

Data Flow

  1. Sources → Multiple Plesk docs sources (5 categories) w/ optional Git auto-cloning
  2. Parsing → Category-specific parsers extract text, title, breadcrumb hierarchy
  3. Chunking → Variable chunk sizes by type (HTML: 3000, PHP: 6000, JS: 5000 chars)
  4. Embedding → BAAI/bge-m3 generates vectors (auto-downloader model on first run)
  5. Storage → LanceDB table w/ UnifiedSchema (vector, text, title, filename, category, breadcrumb)
  6. Retrieval → Query search → rerank → top-5 results w/ relevance scores

Key Files

  • server.py - Main entry point; contains all tools, parsers, indexing logic
  • main.py - Placeholder entrypoint
  • pyproject.toml - Dependencies (fastmcp, lancedb, sentence-transformers)

Development Workflows

Running the Server

source .venv/bin/activate
python server.py

Server initializes on startup: downloads missing sources, builds embeddings, starts MCP server. Models auto-download (~2GB total) on first run.

Adding New Documentation Sources

Edit SOURCES list in server.py. Structure:

{"path": KB_DIR / "category", "cat": "category", "type": "html|php|js", "repo_url": "optional git url"}
  • type determines parser and chunk size
  • repo_url triggers automatic Git cloning if source missing

Read the full file on GitHub · 281 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 281 lines · 2,967 tokens per session scan A cf9788c0a73a

Subscribe to this mod's changes

mcp-plesk-dev-docs copilot-instructions.md is an instructions file published in the GitHub repository barateza/mcp-plesk-dev-docs (2 stars, last pushed 1mo ago), licensed MIT. It adds 2,967 tokens to every session, about $0.0148 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.