The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires several standard Python libraries to function.
It installs google-api-python-client, google-auth-oauthlib, google-auth-httplib2, dateparser, and mcp from official package registries.
[COMMAND_EXECUTION]: The skill executes its own Python scripts to interact with backends.
These scripts are invoked as tools by the agent or through a suggested SessionStart shell hook to provide a daily digest.
[DATA_EXFILTRATION]: The skill synchronizes event context to the user's Google Calendar.
This includes local project paths and session IDs to help users identify the source of a reminder; data is sent exclusively to Google's official and well-known API endpoints.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection from processed data.
Ingestion points: The agent reads event and task data from calendar.py and tasks.py during listing and search operations.
Boundary markers: Content retrieved from external sources (Google Calendar) is not isolated with delimiters or isolation instructions.
Capability inventory: The agent has tools to modify calendar events, manage local task files, and execute shell commands.
Sanitization: No filtering or sanitization is performed on titles or descriptions retrieved from the calendar before they are processed by the LLM.

assistant