Skip to content
SkillModel-led815 tokens in use

replica-build

Rebuilds an app screen by screen from the recon map: shell, vertical slice, then every screen with all its states, written clean-room.

10,140 stars on the repository MIT LicenseJakeschincariol/replica-skill · replica-build
/plugin install replica-skill@replica-skill
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing replica-build costs

101tokens idle
815tokens in use

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

Nothing detected: no shell, network, file writes, git or credentials in its files.

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Good
  • Frontmatter12/15
  • Structure25/25
  • Examples0/15
  • Output10/15
  • Size15/15
  • Resources10/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent

No findings.

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Needs attention
  • Development activity32
  • Maintenance4
  • Community64
  • Maturity12
  • Momentum40

replica-build turns the recon map, architecture plan and design tokens into working screens. It builds the app shell first, then the core flow end to end, then each remaining screen in the order the architecture sets.

Every line is written fresh. The skill never pastes the original’s code, assets or copy, and when it is stuck it goes back to the original as a user, using help articles and your own account.

What You Get

  • An app shell with routing for every screen, layout, tokens wired in and seed data
  • A working vertical slice of the core loop, on a fake data layer if the backend is not built yet
  • Each screen with empty, loading, filled, error, no-permission and long-content states
  • Rows in features.csv set to yes or partial, plus a screenshot per screen for the diff step
  • A replica/build-log.md with one line per screen

Common Use Cases

  • Building a clone screen by screen from a recon map
  • Keeping a feature matrix current as screens are finished
  • Capturing reference-sized screenshots to feed a parity comparison

What Makes It Unique

It carries a per-screen definition of done: every state, 390px and 1440px widths, keyboard-only completion, no console errors and no borrowed copy. It also bans raw hex or pixel values in components, so every value traces to a token.

Under The Hood

The skill is instructions only. The agent writes source code into your project, makes one git commit per screen, updates the feature CSV and build log, and saves screenshots, which means it writes files and runs git and build commands. It tells the agent to read the original app’s help articles and public walkthroughs as a user, but not its code or network calls.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers