mcpField
A
algopack-mega-alerts

algopack-mega-alerts

github.com/moexalgoData & databases440 installsUpdated 19w ago
Repository

About algopack-mega-alerts

A skill for retrieving MOEX AlgoPack Mega Alerts directly from the ISS endpoint and preparing observed anomalies for analysis. It guides data normalization and output formatting while avoiding predictions or recommendations.

Key features

  1. 01Direct Mega Alerts endpoint access
  2. 02Market-wide and ticker-specific queries
  3. 03Normalize ISS columns and data
  4. 04Export tables or charts
  5. 05Treat alerts as observed anomalies

Use cases

  1. 01Fetch alerts for a market or ticker
  2. 02Prepare alert data for tables or charts

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/moexalgo/algopack-skills --skill algopack-mega-alerts

After installing, type/algopack-mega-alertsin Claude Code to confirm it loaded.

Bundled filesreferences/mega-alerts.md

Details

From the project README

ALGOPACK Mega Alerts

Overview

Use this skill for direct Mega Alerts endpoint access. Alerts identify unusual market activity and include historical context in the reference field.

Access note: Promo includes Mega Alerts. Стартовый / Starter free-token availability is planned at T - 1 day; verify current access before promising it.

Quick Start

curl -L "https://apim.moex.com/iss/datashop/algopack/eq/alerts/SBER.json?from=2025-01-01&till=2025-03-30" \
  -H "Authorization: Bearer ${APIKEY}"

Use date=YYYY-MM-DD for market-wide alerts and from=YYYY-MM-DD&till=YYYY-MM-DD for one ticker.

Core Workflow

  1. Choose market eq or fo unless current entitlement proves more coverage.
  2. Fetch alerts and parse the data block.
  3. Normalize ISS columns plus data arrays into rows before filtering, writing tables, or charting.
  4. Group by alert_type, secid, and date/time only when the requested analysis needs grouping.
  5. Parse reference only as historical post-alert context.
  6. Match output to the request: CSV/JSON tables for handoff, simple HTML charts for browser output, Matplotlib or pandas plotting for notebook/script output, and the app's existing charting stack inside an app.
  7. Avoid forecast language.

References

Read references/mega-alerts.md for endpoints, alert types, fields, reference schema, pagination, and output patterns.

Boundary

Do not present alerts as predictions or recommendations. Phrase results as observed anomalies and historical post-alert statistics.

Comments

Markdown supported