The Agent Skills Directory

[SAFE]: The skill instructions specify reading from and writing to the .business/ directory for managing business context and evidence. These operations are aligned with the skill's stated purpose and include safety instructions requiring explicit confirmation before making changes to live messaging or contacting customers.
[PROMPT_INJECTION]: The skill is subject to potential indirect prompt injection as it processes untrusted data from external files such as customer evidence and competitor material.
Ingestion points: SKILL.md instructs the agent to read .business/context.md and other relevant customer/competitor files.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are present.
Capability inventory: The skill utilizes file read and write capabilities within the .business/ directory (SKILL.md).
Sanitization: There is no evidence of sanitization or filtering for the data ingested from external files.

position-business