The Agent Skills Directory

[SAFE]: The skill defines an operational workflow for agent teams, focusing on task coordination and quality assurance through role separation (test-writer, builder, reviewer).\n- [COMMAND_EXECUTION]: The skill configures standard automation hooks that execute local project scripts such as scripts/verify.sh, scripts/fast-verify.sh, and scripts/task-completed.sh. These scripts are used for routine build and test verification.\n- [INDIRECT_PROMPT_INJECTION]: While the agents process external feature documentation, the workflow includes mitigating controls such as automated testing (TDD), file ownership rules, and mandatory reviewer approval to prevent malicious or accidental instruction overrides.\n- [SAFE]: No evidence of obfuscation, data exfiltration, or unauthorized network activity was detected. References to external domains (anthropic.com, schemastore.org) are to trusted or well-known services.

agent-teams