The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or dangerous commands were identified. The skill consists entirely of instructional text focused on secure repository initialization and cloning procedures.
[DATA_EXFILTRATION]: The skill includes specific safeguards against credential leakage, explicitly instructing the agent not to persist secrets in URLs, command history, or repository configurations.
[INDIRECT_PROMPT_INJECTION]: The instructions correctly identify that external data from repositories (such as commit messages or hooks) should be treated as untrusted data unless independently verified, reducing the risk of indirect injection attacks.

setup-repository