The Agent Skills Directory

[SAFE]: The skill instructions are focused entirely on creative writing, strategy, and content optimization. There are no executable scripts, binaries, or commands that interact with the host system in a dangerous manner.
[DATA_EXPOSURE]: The skill reads from a local directory named brand-memory/ to access files like voice-profile.md and audience.md. This is a standard functional requirement for personalizing the agent's output and does not target sensitive system files or credentials.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input in the form of user-provided topics, source content, and past newsletter issues.
Ingestion points: Files in brand-memory/ and user-provided text during the 'Write' and 'Optimize' modes.
Boundary markers: Not explicitly defined in the instructions.
Capability inventory: Limited to text generation, reading specific markdown files, and writing generated newsletter drafts to the local directory.
Sanitization: No specific sanitization or filtering of input content is mentioned.
Risk Assessment: Because the skill lacks network access or high-privilege execution capabilities, the risk of a successful injection leading to significant harm is negligible.

newsletter