skill_skill

Overview

This skill is a modular hub for building, managing, and sharing reusable agent skills, designed to help an AI assistant produce repeatable, verifiable skill deliverables. It is intended for tasks that create or run skill artifacts, orchestrate scripts, and package outputs; primary language is Python and it supports extensible multilingual prompts and assets.

How this skill works

The skill inspects inputs, constraints, and success criteria, then runs parameterized scripts from scripts/ to perform core steps. It pulls optional reference material from references/ and templates from assets/ to assemble final deliverables. Results are validated against the success criteria and a fallback plan is provided if verification fails.

When to use it

• When you need repeatable skill execution or packaging for an AI assistant
• When you want to run or reuse parameterized scripts reliably
• When you must produce verifiable deliverables with clear acceptance criteria
• When you need to integrate references or templates into a skill output
• When you want to share or version reusable agent skills across projects

Best practices

• Clarify inputs, constraints, and explicit success criteria before execution
• Prefer using scripts/ for repeatable or error-prone steps to increase determinism
• Keep destructive operations gated by explicit user confirmation
• Validate outputs automatically against defined acceptance checks and log results
• Store reusable templates and assets in assets/ and reference supporting docs in references/

Example use cases

• Generate and package a new agent skill by running scripts/create_skill.py and validating output against a checklist
• Process input data via scripts/process_input.py, apply templates from assets/, and produce a ready-to-deploy skill bundle
• Run a verification pipeline that checks artifacts, runs unit scripts, and produces a pass/fail report with remediation steps
• Automate non-destructive migrations or updates to skill metadata with explicit confirmation prompts before changes
• Assemble multilingual prompt variants using assets/ templates and include references/ citations for auditability