The Agent Skills Directory

[COMMAND_EXECUTION]: The skill operates by executing shell commands through the gcalcli utility. It specifically utilizes the --iamaexpert flag to bypass interactive safety prompts during event deletion, as documented in SKILL.md.
[PROMPT_INJECTION]: The skill is vulnerable to Indirect Prompt Injection (Category 8). It reads untrusted data from external sources (Google Calendar event titles and descriptions) and possesses write capabilities (add/delete).
Ingestion points: Untrusted data enters the agent context via gcalcli agenda and gcalcli search outputs, as defined in SKILL.md.
Boundary markers: The instructions do not define boundary markers or delimiters to separate tool output from agent instructions.
Capability inventory: The skill has the capability to delete events (gcalcli delete) and create events (gcalcli add and gcalcli import).
Sanitization: There is no evidence of sanitization or filtering of the calendar data before it is processed by the agent.
[COMMAND_EXECUTION]: The skill implements a policy that encourages the agent to skip user confirmation for destructive actions. SKILL.md explicitly states 'Unambiguous actions: execute immediately' for cancel, delete, and edit operations, which increases the impact of potential agent errors or malicious data ingestion.

gcalcli-calendar