ClaudeMap

Birdview

A skill that makes AI coding agents map the architecture and declare the blast radius before they touch a line of code — rendered as one standalone HTML page.

Skills & Commandsskillarchitecturecode-visualizationchange-scopereview

Birdview is a skill for AI coding agents: before editing, the agent maps the project, states which modules and files the task will affect, and records the source evidence behind that plan; it then implements against the map and logs the checks it actually ran. The result is a single interactive HTML page — system map, current change scope, source evidence, full-architecture comparison and verification results on one layout, with light/dark themes, relationship filters, module details and Chinese and English controls — generated from .birdview/architecture.json plus an activity record. No deployed service and no always-on observation: it complements Git diffs, tests and code review rather than replacing them. Install through the third-party skills CLI; the installation guide covers Codex, Claude Code and DeepSeek Harness. v0.2.1, Node 18+, MIT, English and Chinese docs.

What it does

  • Pre-edit architecture map: modules and their ownership, the files the agent plans to touch, and the source evidence behind each claim, all on one page
  • Standalone interactive HTML — light/dark themes, relationship filters, module details, bilingual controls; opens straight from disk, no service
  • Verification record sits next to the plan: the checks the agent actually ran and whether they passed
  • Architecture data and activity records are checked for structure and consistency before the page is generated

When to use it

Anyone reviewing AI-generated changes who wants scope mistakes caught before implementation finishes — especially on large codebases where 'which files will this actually touch?' is the first question to ask.

Getting started

# third-party skills CLI:
npx skills add Qiuner/birdview --skill birdview

# then start a task, e.g.:
#   Use Birdview to show this project's architecture; do not edit code.
# expect .birdview/architecture.json + an HTML map that opens in a browser

Things to know

MIT, v0.2.1. Birdview does not automatically observe every agent action and does not replace Git diffs, tests or code review — the map is only as honest as the agent's declared plan. The screenshot in the README comes from a fictional agent harness bundled in the repository, not from observed production activity. Node 18+; setup guides cover Codex, Claude Code and DeepSeek Harness.

Repository

Qiuner/birdview

Charted

Related on the map

  • Guizang's agent skill turns your repo's recent updates into a finished promo video — real components, code-rendered animation, an original code-written soundtrack — no video editor involved.

    • agent-skill
    • video
    • promo
    • marketing
  • An agent skill that geolocates a photo with nothing readable in it — OSM geometry, elevation skylines and satellite tiles turn “where was this taken?” into a measured answer with an error radius.

    • agent-skill
    • geolocation
    • osint
    • openstreetmap
  • Anthropic's official K-12 teaching skills — lesson planning, differentiation, prep and formative checks, with an eval framework.

    • official
    • anthropic
    • education
    • skills