The Agent Skills Directory

[SAFE]: The skill is primarily instructional and descriptive, outlining a manual approval process for AI agents. It does not include any executable scripts, binary files, or network-active components.
[INDIRECT_PROMPT_INJECTION]: The workflow involves the agent monitoring a local signal file (.loki/signals/CHECKPOINT_APPROVED) to resume execution. This constitutes a potential attack surface for indirect prompt injection if an attacker were to modify the signal files, but the skill itself does not provide the means for such an attack and follows standard workflow practices.

checkpoint-mode