The Agent Skills Directory

[SAFE]: The skill establishes a rigid reporting structure for terminal agent states to prevent vague or ambiguous completion reports. This is a functional constraint that does not introduce security risks.\n- [COMMAND_EXECUTION]: The skill instructs the agent to use standard Bash commands (mkdir, printf, date) to write status logs to a specific project subdirectory (.great_cto/verdicts/). These operations are local, do not require elevated privileges, and are explicitly scoped in the skill's YAML frontmatter.

done-blocked