Enforce explicit Aegis skill loading at turn start.
npx skills add https://github.com/GanyuanRan/Aegis --skill using-aegis
Please help me install this Agent Skill. Skill: using-aegis Source: https://github.com/GanyuanRan/Aegis/tree/main/skills/using-aegis Command: npx skills add https://github.com/GanyuanRan/Aegis --skill using-aegis
Ensures only the explicitly loaded Aegis skill is active at the start of a turn, preventing misrouting and unintended cross-skill activation.
Load only the explicitly relevant Aegis skill at the start of a turn and proceed with normal interaction.
High-intent search queries and answers about installing and using this skill.
To prevent unintended skill activation during turn initialization, you need a hard-priority guard that enforces loading only the explicitly loaded skill. This blocks non-relevant skills and preserves routing context.
Skill routing safety ensures only the explicitly loaded skill is active at the start of a turn. You need routing guardrails when evaluating skill routing across a project baseline to prevent misrouting and cross-skill activation.
To document skill routing decisions for auditing, implement a guard that not only enforces loading the correct skill but also records the routing choice. This creates audit-ready decisions for later review and compliance.
Yes, you can enforce a specific skill to load across a project baseline verification step. A hard-priority guard validates which skill should answer while preserving the current project context during turn initialization.
Cross-skill activation happens when routing decisions fail to isolate the active skill at turn start. You block it by implementing skill gating that enforces loading the correct skill and prevents unintended behavior.
Phife726Discover and load relevant skills before responding to any request.
abnejLLCValidate audit evidence artifacts for completeness, relevance, and authenticity.
cipherstashTest skill resilience against rationalization across RED, GREEN, and REFACTOR phases.
rbergmanCoordinate subagent delegation and review cycles at session start.
