The Agent Skills Directory

[SAFE]: The skill consists entirely of instructional markdown content and documentation regarding development workflows. It does not contain any executable scripts, external download triggers, or network operations.
[INDIRECT_PROMPT_INJECTION]: The skill implements a robust defensive posture by including a dedicated 'Security Model' section. This section explicitly directs the agent to treat all inputs—such as existing code, test logs, and failure messages—as untrusted data rather than authoritative instructions, effectively mitigating potential indirect prompt injection attacks from processed files.
[COMMAND_EXECUTION]: While the instructions mention the execution of shell commands for testing, it correctly delegates the mechanics to project-specific skills and emphasizes using targeted commands scoped to specific tests rather than broad or destructive operations.

tdd