The Agent Skills Directory

[SAFE]: The skill manages internal session state and records brainstorming results by reading from and writing to the .lcs/ directory within the local workspace. These operations are limited to Markdown files and structured state files for project organization.
[DATA_EXFILTRATION]: No evidence of network requests, external data transfers, or hardcoded credentials was found. All data handling is restricted to the local file system.
[PROMPT_INJECTION]: The skill instructions define clear operational boundaries for brainstorming and explicitly prohibit the creation of code or PRDs during this phase, reducing the risk of unintended agent actions.

lcs-explore