The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides a bash script template for searching memory that uses a placeholder (<keywords>) within a shell command (grep). If the agent interpolates raw user input into this command without proper escaping or sanitization, it could lead to local command injection via shell metacharacters (e.g., $(...), `...`, or ;).
[INDIRECT_PROMPT_INJECTION]: The skill is designed to persist user-provided information into MEMORY.md, which is explicitly stated to be loaded at the start of every subsequent session. This creates a mechanism for instructions to persist across sessions, potentially influencing the agent's future behavior if malicious content is saved.
Ingestion points: The agent reads from MEMORY.md during searches and the platform loads it at session start.
Boundary markers: None identified in the storage format; entries are simple markdown list items.
Capability inventory: The skill has file-write access to MEMORY.md and uses grep for file searching.
Sanitization: The instructions do not define any sanitization or validation logic for the content being saved to memory.

remember