The Agent Skills Directory

[SAFE]: The skill acts as a high-level router and does not directly perform file writes, network operations, or command execution. It delegates complex tasks to specialized engines while enforcing that delegated writes stay within existing authorization.
[SAFE]: No evidence of prompt injection or obfuscation was detected. The skill uses disable-model-invocation: true in its configuration, which is a security-hardening measure that prevents the model from spontaneously invoking the skill outside of controlled triggers.
[SAFE]: The 'Spec' mode instructions explicitly promote security best practices, such as requiring independent implementation reviews, mandatory CI/CD gates, and verified acceptance evidence before a feature is marked complete.

prd-dispatch