agent-sort
by affaan-magent-sort is a repo-aware skill for building an evidence-backed ECC install plan. It helps sort skills, commands, rules, hooks, and extras into DAILY vs LIBRARY buckets so you install only what the project actually uses. Use the agent-sort skill for install decisions, agent-sort usage, and a practical agent-sort guide for Skill Authoring workflows.
This skill scores 66/100, which means it is worth listing for directory users who need repo-aware ECC install planning, but it should be adopted with some caution because the repository gives strong intent and workflow structure without bundled automation or deep supporting assets.
- Clear trigger and use case: build an evidence-backed ECC install plan for a specific repo, especially when trimming a full install is needed.
- Operationally explicit workflow: instructs users to sort components into DAILY vs LIBRARY buckets and produce ordered artifacts including an inventory, install plan, and verification report.
- Repo-aware and constraint-driven: requires concrete repository evidence, says not to install unusable hooks/rules/scripts, and avoids generic preference-based decisions.
- No install command or support files are provided, so users must integrate it manually into their workflow.
- The skill is focused on one decision pattern rather than broad automation, so its value is narrower if users want a general-purpose repo analysis skill.
Overview of agent-sort skill
agent-sort is a repo-aware planning skill for turning a broad ECC setup into a smaller, evidence-based install surface. It helps you decide what should load daily versus what should stay searchable in a library, so you do not over-install skills, rules, hooks, or scripts that the project will not use.
This is best for maintainers, agent workflow owners, and anyone doing agent-sort install work on an existing repo. The real job-to-be-done is not “organize files”; it is to produce a defensible install plan from the actual codebase, with fewer guesses and less ECC noise.
What agent-sort is for
Use the agent-sort skill when you need to classify repo-specific ECC pieces into practical buckets, especially in a project with mixed workflows or partial ECC adoption. It is useful when you want daily-loaded surfaces to stay tight while preserving reference material for later lookup.
What makes it different
The agent-sort skill is evidence-driven: the repository content decides the install plan, not personal preference. Its strongest value is reducing accidental installs and surfacing what the project truly uses, which makes it more decision-oriented than a generic prompt.
Best fit and misfit cases
It fits repos where ECC is present but too broad, too repetitive, or drifting away from the project’s real workflow. It is a weaker fit if you want a fast one-off prompt to summarize a repo, or if the project does not use ECC-native surfaces at all.
How to Use agent-sort skill
Install and context setup
Start with the agent-sort install flow in the target repository, then read skills/agent-sort/SKILL.md as your operating guide. The skill is designed to work from repository evidence, so install it alongside the repo you are evaluating, not in isolation.
Inputs that improve output quality
Give the skill a specific target repo, a clear scope, and the decision boundary you care about. Strong input looks like: “Sort this repo’s ECC surfaces into DAILY vs LIBRARY for the current app, preserving only what is needed for routine agent runs.” Weak input is vague, like “organize the project” or “make it cleaner.”
Recommended workflow
Begin by checking the repo’s own README, AGENTS-style instructions, and any ECC-related folders or manifest files. Then use agent-sort usage to map concrete evidence to daily-load candidates, library-only references, and anything that should not be installed at all. The goal is a concise install plan with a verification step, not a speculative re-architecture.
Files to read first
Start with SKILL.md, then inspect the repo tree for README.md, AGENTS.md, metadata.json, and any rules/, resources/, references/, or scripts/ folders. In this repo, SKILL.md is the main source because there are no support files to cross-check.
agent-sort skill FAQ
Is agent-sort only for ECC projects?
Yes, agent-sort is built around ECC-native surfaces and install decisions. If your repository does not use that pattern, ordinary prompting is usually simpler and more appropriate than the agent-sort skill.
How is this different from a normal prompt?
A normal prompt can produce a quick recommendation, but agent-sort pushes for evidence-backed classification and an install plan. That matters when you need repeatable agent-sort usage across a repo instead of a one-off opinion.
Is agent-sort beginner-friendly?
It is beginner-friendly if you can identify the repo you want to evaluate and you are willing to read the source tree first. The main requirement is not deep ECC expertise; it is being clear about what should be installed daily versus kept as library material.
When should I not use agent-sort?
Do not use it if you only need a generic repo summary, if the project has no ECC surfaces to sort, or if you are looking for a broad refactor rather than an install decision. In those cases, the overhead of agent-sort for Skill Authoring is unnecessary.
How to Improve agent-sort skill
Give sharper repo evidence
The best agent-sort results come from concrete repository signals: file paths, current workflow names, and examples of what the agent actually touches. If you can point to specific commands, folders, or rule files, the daily/library split will be more accurate.
State the install boundary clearly
Say whether you are sorting for the whole repo, a subpackage, or a narrow workflow. A precise boundary prevents the skill from over-classifying reference material as daily load, which is a common failure mode in agent-sort install work.
Review the first pass for over-loading
If the output feels too heavy, ask which items are truly required every run and which are only lookup material. The most useful iteration step is trimming daily installs until the plan reflects routine use, not theoretical usefulness.
Preserve useful library material
A strong agent-sort guide does not delete everything that is not daily. Keep borderline items in the library if they support later decisions, because the skill is meant to reduce default load, not erase project knowledge.