agent-sort

Classifies ECC skills, commands, rules, and hooks into DAILY or LIBRARY buckets using repo evidence.

1|Updated Jul 25, 2026
One-click install
npx skills add https://github.com/zacgoodwin/AIBootstrap --skill agent-sort-zacgoodwin
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-sort
Source: https://github.com/zacgoodwin/AIBootstrap/tree/main/.claude/skills/agent-sort
Command: npx skills add https://github.com/zacgoodwin/AIBootstrap --skill agent-sort-zacgoodwin

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Full ECC installs load every skill, rule, and hook regardless of whether a project actually uses them, creating noisy context and stale surfaces. This Skill builds an evidence-backed install plan so a repo only loads what its real stack needs. ## Core Features & Use Cases - Evidence-Based Classification: Sorts agents, skills, commands, rules, hooks, and extras into DAILY (always loaded) or LIBRARY (searchable on demand) buckets, with every DAILY decision citing concrete repo evidence like file extensions, lockfiles, and framework configs. - Parallel Review Passes: Splits the audit into six passes (agents, skills, commands, rules, hooks/scripts, extras) that can run as parallel subagents or sequentially. - Install Plan and Verification: Produces a DAILY/LIBRARY inventory, an actionable install plan, an optional skill-library router, and a verification report confirming the result matches the repo stack. - Use Case: A TypeScript/Next.js repo that inherited the full ECC bundle uses this Skill to demote Python and Django rule sets to LIBRARY, keep frontend and TypeScript surfaces in DAILY, and remove incompatible hooks. ## Quick Start Ask the agent to sort the ECC components in this repository into DAILY and LIBRARY buckets based on the actual stack and produce an install plan.

Frequently Asked Questions about agent-sort

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I trim an ECC install to only what my project needs?

Run a classification pass that sorts every ECC skill, command, rule, hook, and extra into DAILY or LIBRARY buckets based on repo evidence like file extensions, package managers, and framework configs. Then apply the resulting install plan and verify the result matches your stack.

What is the difference between DAILY and LIBRARY buckets?

DAILY components load every session because they match the repo's active language, framework, or workflow. LIBRARY components stay accessible through search or a router skill but are not loaded by default, reducing context overhead without deleting anything.

What repo evidence is used to classify ECC components?

Classification uses file extensions, package managers and lockfiles, framework configs, CI and hook configs, build and test scripts, dependency manifests, and repo docs. Commands like rg --files and reading package.json, pyproject.toml, Cargo.toml, or go.mod establish the real stack.

Can the review passes run in parallel with subagents?

Yes, the review splits into six passes covering agents, skills, commands, rules, hooks and scripts, and extras, which can run as parallel subagents. If subagents are unavailable, the same passes run sequentially with identical results.

When should I not use evidence-based ECC sorting?

Avoid it when a repo genuinely needs the full ECC surface loaded every session, or when the stack is still undecided and classification would be guesswork. It is designed for repos with a clear, established toolchain.