ashe-li

ashe-li/agent-skills

Agent Claude Code

A task-sizing step that labels work as small, medium, or needing multiple sessions before detailed planning begins.

ashe-li/agent-skills

Agent Claude Code

A document and plan review agent that checks whether requirements are covered, the work is practical, supported by industry or community evidence, and safe.

ashe-li/agent-skills

Agent Claude Code

A documentation-updating agent that scans recent code changes, or a specified range of changes, and updates affected files such as README files, documentation, and code maps. It returns a change manifest for review.

ashe-li/agent-skills

Agent Claude Code

A guide to test-driven development, or TDD: writing a test that fails, adding code to make it pass, and then improving the code. It is intended to be used before new features or bug fixes and includes an 80% code-coverage threshold.

ashe-li/agent-skills

Skill Claude Code

A development-planning skill that inventories available resources and uses a planning agent to create a detailed implementation plan as a Markdown file in plans/active/.

ashe-li/agent-skills

Skill Claude Code

An independent process for checking evidence behind a technical decision or practice across research, industry standards, best practices, and community opinions, including opposing views.

ashe-li/agent-skills

Skill Claude Code

A fact-checking gate for reports produced by coding workflows such as pull-request and security reviews.

ashe-li/agent-skills

Skill Claude Code

A visual checking process for comparing Figma designs with local web implementations. Figma is a tool for creating interface designs, while a screenshot diff compares images to find visual changes.

ashe-li/agent-skills

Skill Claude Code

A skill that creates a self-contained handoff prompt summarizing a conversation’s goal, progress, decisions, and unfinished work.

ashe-li/agent-skills

Skill Claude Code

A workflow that reads requirements from a Notion page and connects them to a design implementation plan. Notion is a collaborative app for storing documents and project information.

ashe-li/agent-skills

Skill Claude Code

A procedure for moving completed plans from an active folder to a completed folder, while recording verification results and the completion time.

ashe-li/agent-skills

Skill Claude Code

A Python state machine that follows the dependency graph in a plan.md file. A dependency graph shows which tasks must be completed before others can start.

ashe-li/agent-skills

Skill Claude Code

A pull-request acceptance workflow that checks a preview, captures screenshots, visually reviews them, and posts pass-or-fail evidence as a comment.

ashe-li/agent-skills

Skill Claude Code

A workflow for finishing repository work and opening or updating a pull request. A pull request is a proposed code change that others can review before it is merged.

ashe-li/agent-skills

Skill Claude Code

A knowledge-base update workflow described in Chinese. It runs document updates, reviews, context organisation, and pattern extraction in sequence so changes from a work session can be recorded.

ashe-li/agent-skills

Skill Claude Code

A skill for managing Git worktrees throughout their lifecycle. A worktree is a separate working folder connected to the same Git repository.

ashe-li/agent-skills

Skill Claude Code

A workflow for coordinating several coding agents through a multi-step plan. It assigns each step, checks the results, reports status, and applies a knowledge check before finishing.

ashe-li/agent-skills

Skill Claude Code

A reporting assistant that writes work results or progress updates to Notion, a workspace for shared notes and project information.

ashe-li/agent-skills

Skill Claude Code

A release pull-request assistant that creates a title and description from the commits between two Git branches and writes release notes in Traditional Chinese.

ashe-li/agent-skills

Skill Claude Code

A root-cause analysis workflow for investigating an alert, error report, or failed build. It gathers read-only evidence from monitoring, error tracking, history, documentation, and infrastructure, then combines it into a timeline and proposed next steps.

ashe-li/agent-skills

Skill Claude Code

An alert investigation workflow for a firing Grafana or Loki alert, using live Grafana and Kubernetes data to find the likely root cause. It classifies the alert as a harmless false alarm or a real regression, then produces a Traditional Chinese root-cause report.