Calendar

Name/description match the actual behavior (gcalcli CLI calls). Requiring the gcalcli binary and providing a pip install for gcalcli is proportionate to the stated purpose. However, the SKILL.md mentions needing GOOGLE_CALENDAR_API_KEY or CALDAV_URL/CALDAV_USER/CALDAV_PASS while the skill metadata lists no required environment variables or credentials — this mismatch is unexpected.

SKILL.md contains concrete gcalcli commands (agenda, add, delete) which stay within calendar management. But it explicitly refers to API keys/CALDAV credentials as required inputs; those credentials are sensitive and the instructions do not say how they are supplied or stored. The instructions give broad examples but do not constrain where credentials come from, which increases the risk if an agent is given access to environment/config.

Install is a straightforward pip install gcalcli which is expected for this tool. pip installs are common and reasonable here, but they carry the usual supply-chain risk of installing packages from PyPI. No obscure or remote download URLs are used.

The skill metadata declares no required environment variables, yet SKILL.md explicitly references GOOGLE_CALENDAR_API_KEY or CALDAV_URL/CALDAV_USER/CALDAV_PASS. Sensitive variables (API keys, usernames, passwords) are implied but not declared as required — this is a red flag because the agent or user might be asked to provide or expose secrets without clear metadata or provenance.

The skill does not request always:true and does not claim to modify other skills or global agent config. Agent autonomous invocation is allowed (the platform default) but that by itself is expected for typical skills.