The Agent Skills Directory

[SAFE]: The skill implements a mandatory "Evidence Discipline" policy which requires the classification of all material claims as fact, inference, assumption, or unknown, preventing the fabrication of metrics or research data.
[SAFE]: Multi-agent coordination includes a "marketing-skeptic" agent to serve as a counterweight, ensuring that strategic recommendations are challenged and verified rather than accepted as automatic authority.
[SAFE]: Operational integrity is maintained through specific completion gates that require material uncertainty to be explicitly stated and reversal evidence to be defined before a decision is approved.
[SAFE]: Routing mechanisms rely on local script references (dynamic_router.py, neural_router.py) for managing complex task dependencies within the project's own repository structure, rather than fetching external or untrusted code.

ai-discovery-strategy