This one does exactly what it says: it makes Claude pause and ask clarifying questions before jumping into implementation when a request is vague or could mean multiple things. The workflow is straightforward: check if key details like objective, scope, or constraints are missing, then ask the minimum questions needed to avoid building the wrong thing. It explicitly tells you not to use it for clear requests or when you can just peek at the code to figure out what's missing. Honestly, this feels like table stakes behavior you'd want by default, but making it explicit as a skill means you can toggle it based on your working style. Comes from Trail of Bits originally, now maintained by Hypergiant, and has passed security audits from Gen Agent Trust Hub, Socket, and Snyk.
npx -y skills add gohypergiant/agent-skills --skill ask-questions-if-underspecified --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
sickn33/antigravity-awesome-skills
moizibnyousaf/ai-agent-skills
github/awesome-copilot