mcp-layer

A set of three outward-facing MCP tools for finding event travel packages, viewing trip details, and preparing checkout links. MCP is a way for AI agents to call external tools through a defined interface.

In plain words
What is it for?
Searching trips by topics and travel constraints, retrieving package details, and creating a checklist of checkout links.
Why use it?
It gives agents a consistent, validated interface for the travel service and returns structured results that clients can parse reliably.

Cursor rule for Cursor

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 rules/evilfreelancer/zaezd/mcp-layer
Clone the repo
git clone --depth 1 https://github.com/EvilFreelancer/zaezd

Made for: Cursor.

Per session 0 Nothing until a file matches its globs; then the whole rule loads.
When invoked 1,260 The whole file, excluding the scripts and references it only reads on demand.
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 $0.00000 $0.01260
Opus 5 $0.00000 $0.00630
Sonnet 5 $0.00000 $0.00252
Haiku 4.5 $0.00000 $0.00126

Measured 2d ago against content hash 2f5820fce78f, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

mcp-layer 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.

.cursor/rules/mcp-layer.mdc · 102 lines

How it starts

The opening of the file, as written. The whole thing — 102 lines — stays where its author put it; the contents beside it link to each section on GitHub.

MCP layer: three tools and one widget

src/mcp/ exposes the product to agents. It is an adapter: it validates input, calls the orchestrator, and shapes the answer. No business logic lives here.

The three tools

Tool Arguments Returns
find_event_trips topics[], origin, budget?, date_from?, date_to?, adults? the event, up to three packages, the coverage note, a link to the web version
get_trip_details trip_id, package package details, hotel, rates, weather when available
create_trip_checkout trip_id, package a checklist of two or three links, each carrying its actual kind

These exact names appear in the code, the README, the user guide and the architecture diagram. Renaming one means renaming it everywhere in the same change.

  • All three declare outputSchema and return structuredContent. This is precisely the gap measured in Tutu MCP, and closing it removes a whole class of client parse errors.
  • Annotations are honest: readOnlyHint: true and destructiveHint: false on all three, idempotentHint: false on create_trip_checkout because the links expire.
  • Input validation is forgiving (see @data-sources.mdc): arrays accepted as array, JSON string or comma-separated string; numeric strings coerced; a missing argument object treated as empty. Reject nothing that can be understood.

What this layer must not do

  • Proxy raw Tutu tools outward. The point of the gateway is a product contract of three verbs, not a re-export of sixteen searches.
  • Keep trip state on the server. trip_id is a compact encoding of the request, not a key in a store, and /t/:id is reproducible from it alone.
  • Create a cart. The user opens the link and the cart appears in their own Tutu session. That is Tutu's legal model and it is not ours to work around.
  • Store a checkout link in a snapshot. Links are rebuilt live on click; if the live call fails, fall back to search_results_url with an honest label.

Read the full file on GitHub · 102 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 · 102 lines · 0 tokens per session scan A 2f5820fce78f

Subscribe to this mod's changes

mcp-layer is a cursor rule published in the GitHub repository EvilFreelancer/zaezd (1 stars, last pushed 13d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,260 tokens. 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.