The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands (Grep) using keywords extracted directly from user conversations. The instructions in SKILL.md and references/protocol.md do not include sanitization requirements or safety boundaries (such as shell escaping or the use of -- to separate flags from arguments), creating a potential command injection surface if the agent obeys a malicious user-provided string.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) as it processes untrusted user data to determine file write operations and shell command parameters.
Ingestion points: User session input is parsed for events, decisions, and preferences (documented in SKILL.md).
Boundary markers: None identified. There are no instructions to the agent to ignore or escape instructions embedded within the user's "memories".
Capability inventory: The skill performs file system writes to the AI_MEMORY/ directory and is instructed to execute shell tools (Grep).
Sanitization: No sanitization or validation logic is present in the instructions or the provided scripts/update_memory.py helper script.

mem-record