The Agent Skills Directory

[COMMAND_EXECUTION] (HIGH): The skill requests the Bash and Write tools to manage task states and log metrics like task_completion_streaks. Because it ingests untrusted user data (task names and descriptions) and has execution capabilities, it presents a high risk for Indirect Prompt Injection. If an attacker provides a 'task' containing malicious shell syntax, the agent might inadvertently execute it while attempting to log or track the task.
Ingestion points: User task descriptions, session content, and focus assistance requests in SKILL.md.
Boundary markers: None present; the skill instructions do not define delimiters to separate user data from system commands.
Capability inventory: Bash, Write, and Read tools specified in allowed-tools.
Sanitization: None present; there are no instructions to escape or validate user-provided strings before processing them through system tools.

adhd-productivity