The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were identified in the skill. The instructions prioritize user safety by enforcing explicit account selection and recommending confirmation steps before performing any write or delete operations.
[INDIRECT_PROMPT_INJECTION]: The skill interacts with external data from Google Classroom (such as announcements and submissions), which is a potential vector for indirect prompt injection. However, the skill explicitly mitigates this by instructing the use of the --wrap-untrusted flag and boundary markers for agent-readable content.
Ingestion points: Data retrieved from Google Classroom via subcommands like announcements, courses, and submissions.
Boundary markers: The skill recommends using --wrap-untrusted to delimit external data.
Capability inventory: The tool allows for data retrieval, modification, and deletion within the Google Classroom environment.
Sanitization: Relies on the provided CLI tool's handling and the agent's application of the recommended safety flags.

gog-classroom