The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes Git and shell commands to restore files to a previous state. This is the primary function of the skill and is handled with appropriate safety measures such as previews and safety backups.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied checkpoint tags within shell commands, creating a potential injection surface. 1. Ingestion points: The $CHECKPOINT_TAG parameter in SKILL.md. 2. Boundary markers: Includes a mandatory manual confirmation step where the user must type 'ROLLBACK'. 3. Capability inventory: Shell command execution via git, cp, and echo for file restoration and logging (SKILL.md). 4. Sanitization: Includes a validation step using git tag -l to ensure the tag exists before execution.

skill-rollback