Design Claude Code agents with experiential identities and structured prompts.
npx skills add https://github.com/101mare/skill-library --skill agent-builder-101mare
Please help me install this Agent Skill. Skill: agent-builder Source: https://github.com/101mare/skill-library/tree/main/skills/meta/agent-builder Command: npx skills add https://github.com/101mare/skill-library --skill agent-builder-101mare
This Skill addresses the challenge of creating AI agents that are not only functional but also highly effective and specialized, moving beyond generic instructions to deeply ingrained, research-backed identities.
Use the agent-builder skill to learn how to create a new agent for analyzing user feedback.
High-intent search queries and answers about installing and using this skill.
Design Claude Code agents by crafting prompts that leverage specific experiential identities, using structured frontmatter fields and multi-file organization to achieve 10-60% improved accuracy over generic expert labels.
Structure multi-file agent prompts by organizing complex agents into separate files and applying consolidation patterns to merge specialized agents into more powerful, consolidated ones with optimized tool selection.
Generic responses occur when agents lack specific experiential identities. Fix this by implementing research-backed identity design with anti-pattern avoidance sections defining what the agent refuses to do for reliable specialization.
Consolidate multiple specialized agents by applying structured consolidation patterns that merge distinct experiential identities into a unified agent while maintaining specialized behavior through carefully designed frontmatter fields.
Prompt engineering with experiential identities improves accuracy by 10-60% compared to generic expert labels, because deeply ingrained research-backed identities provide more insightful and accurate agent behavior for specialized tasks.
Add an anti-pattern section when building reliable agents that require strict behavioral boundaries, defining what the agent refuses to do ensures consistent specialized behavior beyond what generic instructions provide.
catlog22Validates delegation prompts against agent role definitions for conflicts and missing contracts.
lucasmiachon-blipDefine, test, and iterate Claude Skills with frontmatter-driven workflows.
ViewWayProvide expert instruction for crafting effective AI prompts.
mathur-exeRefine prompts to steer AI behavior toward coherent extrapolated volition.
