The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes the platform-provided ai-native-backend-contract and the external grilling skill to manage the interview workflow and state changes. These operations are core to the skill's primary purpose and do not involve arbitrary command execution.
[PROMPT_INJECTION]: The skill ingests external content from 'target work' records and repository documentation. While this creates an indirect prompt injection surface, the risk is mitigated by explicit boundaries that prevent the skill from modifying local files and require human confirmation for decisions.
[SAFE]: Accesses the .and/config.yml file to retrieve configuration settings for the backend connection. This is a standard and expected procedure for environment setup.

issue-grill