The Agent Skills Directory

[COMMAND_EXECUTION]: The skill's template includes standard verification commands such as bun run lint, bun run typecheck, and bun test. These are intended for verifying code changes in a local development environment and do not constitute malicious or hidden execution.
[PROMPT_INJECTION]: The skill ingests untrusted data from user arguments and external 'epic' files during the planning process. While this creates a surface for indirect prompt injection, the risk is mitigated by the mandatory use of ExitPlanMode, which requires the user to explicitly review and confirm the plan before any implementation or further actions are taken.

agile-story