agent-skills_skill

This skill guides you in creating and auditing high-impact Agent Skills, enabling concise prompts, modular design, and efficient context management.
  • Shell

7

GitHub Stars

1

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill gwenwindflower/.charmschool --skill agent-skills

  • SKILL.md20.1 KB

Overview

This skill guides creators through building and improving compact, reliable Agent Skills that extend an AI agent with focused workflows, domain knowledge, and reusable resources. It emphasizes token-efficient design, clear trigger metadata, and a small set of bundled resources (scripts, references, assets) to keep the agent performant and predictable.

How this skill works

The guide teaches how to design skills with three progressive loading layers: minimal metadata always available, the core instructions loaded only when the skill is invoked, and supplemental resources loaded on demand. It recommends splitting content so the agent only reads what it needs and bundling deterministic code or large assets outside the immediate context.

When to use it

  • Creating a new Agent Skill from scratch
  • Auditing and tightening an existing Skill for token efficiency
  • Deciding which code, docs, or assets to bundle with a Skill
  • Designing triggers and metadata to ensure correct skill invocation
  • Refactoring a skill that has become too large or redundant

Best practices

  • Favor concise, essential instructions; assume the agent already has general knowledge
  • Design explicit trigger metadata that describes when the skill should run
  • Use low degrees of freedom for fragile operations; increase flexibility only after stability
  • Bundle deterministic scripts for repeatable operations and large assets outside core context
  • Keep long reference material in separate files and link to them so they load only when needed
  • Avoid including extra documentation or unrelated files that bloat the skill

Example use cases

  • Create a PDF tooling skill with a bundled rotation script and a short workflow summary
  • Audit an image-processing skill to move large format examples into separate reference files
  • Build a database query skill that stores table schemas in references and exposes a concise query pattern
  • Refactor a multi-provider deploy skill to present a single selection step, then load provider-specific references
  • Initialize a reusable frontend template asset and a script to scaffold projects deterministically

FAQ

Keep core instructions minimal and procedural. Include only what the agent needs to perform the task; move long examples, schemas, or domain docs to separate reference files.

When should I include executable scripts in the skill?

Include scripts when tasks require deterministic, repeatable code or when the same code is being rewritten frequently. Scripts reduce token use and improve reliability.

How do I decide freedom levels for a workflow?

Match specificity to fragility: use strict, step-by-step rules for delicate operations and higher-level guidance where multiple valid approaches exist.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational