Skill Building

Route skill creation and editing requests to the right specialized skill creator

Create, edit and benchmark agent skills with evals and variance analysis

Turn SOPs, links or stated workflows into installed agent skill packages

Generate agent skills from a web page, tutorial or video with extracted steps and scripts

Profile models, compile skills ahead of time and run skill benchmarks with skvm

Send failed or confusing skill runs to skvm for reviewable optimization proposals

Find unused K-skills by checking trigger counts and recommend removals

Install the K-skill bundle, set credentials and verify the CLI runtime

Move approved skills into a parent bundle with timestamped folder backups

Recommend skill groups to consolidate into bundles in a read-only review

Group a large skill inventory into parent bundles with hierarchical sub-skills

Loop review and fixes on an agent skill until the reviewer comes back clean

Generate agent skills from the Trail of Bits Testing Handbook pages

Extract reusable architecture patterns from a codebase and turn them into skills

Turn source code into an educational skill with architecture notes and examples

Search local skills and browse cloud packages to reuse proven workflows

Generate SKILL.md files from extracted architecture patterns

Build GSD extensions that add tools, slash commands and event hooks

Write and refine SKILL.md files with clear structure, triggers and examples

Turn findings from a finished spike into a reusable project skill

Shape raw content or an idea into a compliant agent skill through guided conversation

Turn raw PM content into a compliant, publish-ready repo skill without breaking standards

Guide creating or updating agent skills with clear triggers and workflows

Evaluate a local Codex plugin, explain its score and what to fix first

Audit a local agent skill, explain its score and rank what to fix first

Turn plugin evaluation findings into a rewrite brief for a Codex skill

Design custom evaluation metric packs that score local plugins and skills

Score a local skill or plugin, measure token use and benchmark starter scenarios

Create and edit SKILL.md skills, including their structure and bundled files

Audit installed agent skills for context budget, duplicates and vague descriptions

Design and build high-quality agent skills through a structured conversation

Convert an article or prompt pack into focused, reusable agent skills

Turn a web design technique you built into a reusable skill with a demo

Scaffold Hana plugins with manifests, runtime tools, WebView UI and SDK templates

Create, extract or update a skill from the current session or existing guidance

Score and improve SKILL.md files with a 9-point rubric, keeping only validated changes

Write concise reusable skills with valid triggers and testable completion criteria

Create, validate and package original skills, optionally installing them

Turn website exploration into reusable scraper skills with scripts

Audit agents, skills and commands for quality and production readiness scores

Audit agent skills for metadata, trigger accuracy and completion criteria before release

Coach you to pick an agent, command, skill, hook or rule and generate it

Scaffold a Claude Code skill with SKILL.md, frontmatter and bundled resources

Explain four-layer skill composition, skill marketplaces and a self-learning loop

Vet third-party agent skills for red flags, permission scope and suspicious patterns

Install the CCG bundle of quality gates, doc generator and orchestration skills

Create and repair agent skills, sharpen trigger descriptions and tidy skill structure

Apply quality review improvement plans to fix and update existing agent skills

Evaluate skill descriptions, structure and writing quality and produce a report