Skip to main content
Rewrite Skills are reusable capability packs for AI agents. They define focused workflows, reusable instructions, and implementation patterns so teams can ship faster with predictable quality.

What you get with Skills

Repeatable workflows

Turn common engineering tasks into reusable playbooks instead of rewriting prompts every time.

Team consistency

Keep execution style, coding standards, and decision criteria aligned across contributors.

Faster execution

Reduce context-switch cost with structured steps and built-in task boundaries.

Easier onboarding

New teammates can start with proven flows instead of inventing process from scratch.

Rewrite Skills repository

RewriteToday/skills

Official skill library with practical templates and workflows you can adapt to your stack.

Install and use Skills

Use the command below to add a skill source:
Example with Rewrite Skills:
After adding, follow your agent environment docs to list, activate, and run installed skills.

Where Skills fit in your stack

Planning

Break down large goals into clear, actionable steps.

Implementation

Execute changes with stable patterns for edits, tests, and validation.

Operations

Handle triage, CI fixes, and documentation workflows with less variance.

Typical skill structure

Define exactly what the skill owns and when it should be triggered.
List the sequence of actions: discovery, implementation, validation, and handoff.
Add non-negotiables such as safety checks, style constraints, and prohibited operations.
Include scripts, templates, and references to reduce manual repetition.

Best practices

  • Keep each skill narrowly scoped to one outcome.
  • Prefer concrete steps over broad generic instructions.
  • Add validation criteria so success is measurable.
  • Version and evolve skills as your codebase and tooling change.
  • Link each skill to real workflows used by your team.

API Introduction

Reference for endpoints, auth, and operational constraints used by AI workflows.

Ecosystem

See how Rewrite packages can support agent-driven implementations.