The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documentation explicitly states that it creates a recurring cron job for daily gift delivery. This establishes a persistence mechanism that allows the skill to execute actions automatically across sessions. Additionally, the skill utilizes shell scripts located in a scripts/ directory to perform rendering and network-based asset fetching.
[DATA_EXFILTRATION]: The skill is designed to read and synthesize content from conversation context, memory files, and personal 'emotion peaks.' This data is then sent to external third-party APIs for image and video generation. This pattern represents a risk where sensitive user information stored in the agent's memory could be transmitted to external servers without specific sanitization.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its core functionality of ingesting and processing untrusted user data from conversation history to influence its 'Editorial Judgment' and 'Creative Concept' stages.
Ingestion points: Conversation history, memory files, and relationship-aware context signals.
Boundary markers: None identified; the skill processes natural language context directly from the agent's workspace.
Capability inventory: The skill possesses the capability to execute shell scripts, perform network requests to external APIs, write to the local workspace (workspace/daily-gift/), and manage persistent cron jobs.
Sanitization: No mechanisms for filtering, escaping, or validating the ingested conversation content are described.

daily-gift