Claude Skill MartBrowse skillsQuick linesLearn by videoTerminal guideWhat is a Skill?
 
← Back to list

Agentic TDD (against the running app)

Runs a real red-green-refactor loop against the running app for UI and full-stack behaviour that unit tests cannot express.

Dev & CodingIntermediate★ 320⑂ 61AI score 9/10Last updated: Aug 18, 2026

What it does

It lifts TDD one level up: instead of testing a unit, you declare the consequence that matters on the live app — "clicking Deploy posts to /api/deploy, moves the store to deploying, shows the banner, logs no console errors". Because reticle_act_and_wait({ ref, action, until }) takes the expectation as an argument, the oracle is named before the action runs, so the agent cannot retrofit an assertion around whatever it just built.

  • RED: declare the predicate and require verified: "no"; unknown is not a red
  • GREEN: smallest change, then re-run the same call, unchanged
  • REFACTOR: predicates bind to behaviour (request, signal, state path), not markup
  • Persist with reticle_flow_save / reticle_flow_verify for a free regression suite

It also ranks assertion strength (app-fired signal > state > network > element > absence of console errors) and enforces an honesty rule: never weaken a check to turn a verdict green.

Who it's for

  • Web/full-stack developers using Claude Code who want verification against the real app, not mocks
  • Teams tired of the "write code and hope" agent pattern
  • Anyone who wants an E2E regression suite they never hand-wrote

Examples

  1. Building a Deploy button: assert POST 200 + deploy:started signal + banner + no console errors first, confirm red, then implement.
  2. When the verdict is unknown/unsettled, re-assert with reticle_assert instead of re-driving a side effect.
  3. Refactoring components: re-run the untouched predicate to prove behaviour survived the rewrite.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/reticlehq/reticle/HEAD/skills/agentic-tdd/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the skills/agentic-tdd folder from the GitHub repo reticlehq/reticle into my ~/.claude/skills/agentic-tdd/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/reticlehq/reticle.git /tmp/reticle && mkdir -p ~/.claude/skills && cp -r /tmp/reticle/skills/agentic-tdd ~/.claude/skills/agentic-tdd

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal.
  2. Clone the repo and copy the skill: git clone https://github.com/reticlehq/reticle.git /tmp/reticle && mkdir -p ~/.claude/skills && cp -r /tmp/reticle/skills/agentic-tdd ~/.claude/skills/agentic-tdd
  3. Initialise the Reticle server: RETICLE_INSTALL_SOURCE=npx_skill npx @reticlehq/server@latest init
  4. Follow the install-and-verify skill in the same repo to confirm the MCP connection works.
  5. Restart Claude Code and check that agentic-tdd appears in your skill list.
  6. Start your dev server, then ask Claude to "build this feature test-first" to trigger the skill.
View source on GitHub ↗License: NOASSERTION