A Model Context Protocol server that tracks flight prices over time and helps you decide when to buy. It builds your own price history by saving a snapshot of the price every time you check a route, then uses analysis tools to tell you whether the current price is good or likely to drop.
Flight Price Tracker has clear, well-structured tool definitions with complete schemas and descriptions. All 5 tools follow verb_noun naming (add_route, list_routes, check_price, price_history, analyze_route). Descriptions are substantive (100-250 chars) and explain WHAT, WHEN, and WHY. Input schemas are fully typed with parameter descriptions. However, output schemas are not formally documented, responses are free-text strings rather than structured objects. Error handling is present but generic (e.g., 'Route not found' without recovery guidance). No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk profiles (add_route and check_price are WRITE operations). Parameter descriptions are good but lack explicit constraints (e.g., date format is stated but not validated with regex or enum). Overall solid for a single-domain tool, but lacks production-grade polish around output structure and error recovery.
Register a flight route to track.
Analyze the saved price history for a route: min, max, average, current price, its percentile within history, the recent trend, and a transparent buy/wait verdict.
Fetch the current cheapest price for a tracked route and save a snapshot to the local history. Run this regularly (or on a schedule) to build up data.
List all tracked routes with how many price snapshots each has.
Return the full saved price history for a route as a chronological list.
Output schemas not formally documented. All tools return free-text strings instead of structured JSON objects. LLMs cannot reliably parse unstructured responses or chain results to downstream tools.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). add_route and check_price are WRITE operations that modify state, but this is not declared in the schema. Agents cannot distinguish safe reads from risky mutations.
Error messages lack recovery guidance. 'Route not found' does not suggest calling list_routes() first. 'Invalid date format' does not show the expected YYYY-MM-DD pattern. Errors should guide the LLM to the next action.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 58 | 2026-07-28+ | v2 |
Parameter constraints not formally declared. Date parameters accept strings but lack regex pattern or format constraint. return_date is optional but no enum or boolean flag clarifies one-way vs round-trip intent.
No pagination or result limits documented. list_routes and price_history could return unbounded lists. Large result sets waste tokens and risk context window exhaustion.