The Agent Skills Directory

[SAFE]: The skill implements read-only access to the Yandex Metrica API, limiting its scope to traffic and acquisition reporting (country, source, and UTM parameters).
[SAFE]: Secure credential management is observed. The skill retrieves its OAuth token from the environment variable $YANDEXMETRICA_ACCESS_TOKEN and includes a safeguard in scripts/yandex_metrica.py to redact this token from any API error messages before they are displayed or logged.
[SAFE]: Input validation is performed on critical parameters. The report function in the Python script ensures that counter_id is a positive numeric value and verifies that the start date does not occur after the end date, preventing malformed requests.
[SAFE]: The skill exclusively communicates with official Yandex API endpoints (https://api-metrika.yandex.net). The domain is a well-known service infrastructure, and the network operations are limited to these verified endpoints.
[SAFE]: The test suite (tests/test_yandex_metrica.py) uses standard mocking and dynamic module loading to verify that dimensions and attribution models are correctly mapped to the Yandex API schema.

yandex-metrica