What problem does it solve? AI agents often fail to use available skills correctly: they read skill files directly instead of invoking them, subagents redundantly re-dispatch tasks, and model-invocation restrictions cause confusing errors. This Skill sets the ground rules at the start of every conversation so skills are found and invoked properly. ## Core Features & Use Cases - Platform-Aware Skill Access: Instructs agents how to invoke skills on Claude Code (Skill tool), Antigravity/Gemini CLI (activate_skill), or other environments. - Subagent Guardrails: Detects subagent contexts and prevents wasteful nested dispatching of further agents. - Routing Transparency: Requires stating the rationale when choosing between ambiguous skill candidates before invoking one. - Use Case: When starting a new session in the Sixth Sense plugin, this Skill ensures the agent knows how to discover other skills, respects disable-model-invocation flags, and avoids redundant subagent fan-out. ## Quick Start Start a new conversation and let the session hook load this skill so the agent knows how to find and invoke other skills correctly.
