The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs file and directory operations within the ./writing-workspace/materials/ path. It creates directories and reads/writes JSON and Markdown files to maintain a local database of extracted content. These actions are aligned with the skill's primary purpose of information management.
[PROMPT_INJECTION]: The skill processes untrusted external articles which introduces a surface for indirect prompt injection. However, the instructions direct the agent to treat the content as a data source for extraction rather than executing commands or overriding its own logic. This is a common and expected risk for information processing skills.
[SAFE]: There are no network operations, hardcoded credentials, or obfuscated segments. The skill does not download external scripts or execute dynamic code. The logic is transparently defined in the markdown instructions.

material-ingest