The Agent Skills Directory

[SAFE]: The skill is composed entirely of markdown documentation and instructions for the agent to follow when summarizing internal state. It does not include any scripts, binary files, or executable logic.
[SAFE]: No network operations, external downloads, or remote code execution patterns were identified. All operations involve local text processing and writing to standard project-specific paths like .cavekit/.
[SAFE]: The protocol explicitly excludes security-sensitive content, such as authentication details, secrets, and security warnings, from high-intensity compression to ensure they remain human-readable and accurate.
[SAFE]: The skill uses local file access to read session budget information from .cavekit/token-ledger.json, which is a standard practice for managing state in the documented environment.

caveman-internal