agent-ready

Tag a card.Wake up to a pull request.

Connect Trello, pick your model, choose when it runs, and get the PR link in Telegram. Everything runs in GitHub Actions inside your own repo.

setup wizard

Works with the tools you already use

Trello
Anthropic
OpenAI
Gemini
GitHubGitHub
TelegramTelegram

Workflow

From card to PR in minutes.

Minimal setup. Clear handoff. Your repo stays in charge the whole time.

  1. Step 1

    Label it

    Tag any Trello card "agent-ready" in your To Do list.

  2. Step 2

    Agent wakes up

    Runs on your cron schedule or manually from the GitHub Actions tab.

  3. Step 3

    AI reads & codes

    Reads the card, identifies relevant files, and writes the changes.

  4. Step 4

    PR opens

    New agent/ branch, PR with a one-line summary, ready for review.

  5. Step 5

    You get pinged

    Telegram lands with the PR link. Merge it. Card moves to Done.

Features

What you actually get.

5-minute setup wizard

Interactive CLI guides you through every config step. No YAML hand-editing.

Runs on your repo

GitHub Actions on your own infrastructure. Your keys stay yours. Nothing centralized.

Multiple tasks per run

Every labeled card is processed in one run. Each gets its own branch and PR.

Smart failure handling

Failed tasks move back to To Do with error details in Telegram. Other tasks continue.

Timezone-aware scheduling

Pick your timezone. Choose one to four run times. The wizard handles the UTC cron for you.

Multi-AI support

Choose Claude, GPT-4o, or Gemini. Swap anytime. Bring your own key. No lock-in.

Quickstart

Up and running in 5 minutes.

quickstart
Step 1

Run the wizard

$ npx @elmiristic/agent-ready init
Step 2

Follow the prompts

  • Connect your Trello board
  • Add your AI provider API key
  • Configure Telegram notifications
  • Set your timezone and daily run times
  • GitHub secrets are injected automatically
Step 3

Commit and push

$ git add .github && git commit -m "feat: add agent-ready" && git push
Step 4

Label a card

Tag any Trello card "agent-ready" and watch the agent do the rest.

Stack

Keep the workflow. Swap in the agent.

agent-ready sits between the tools your team already uses. No new planner. No new review surface. No new place to chase status.

Existing stack, tighter loop

Start in Trello. Pick the model. Review in GitHub. Get the link in Telegram.

The handoff is clean because every step already fits the way engineering teams work today.

Trello stays the intake point.
Claude, GPT, and Gemini stay swappable.
GitHub Actions keeps execution in your repo.
Telegram closes the loop with the PR link.

Step 01

Start where work already lives

No new queue to manage. The trigger stays in the tool your team already checks all day.

Trello

Tag a card and the run starts from there.

Step 02

Pick the model for the repo

Claude, GPT, and Gemini are all on the table. Change the model without changing the workflow.

Anthropic

Claude is available as a first-class option.

OpenAI

Use GPT when your team is already on OpenAI.

Gemini

Swap to Gemini without changing the rest of the flow.

Step 03

Keep execution and delivery close

The run happens in GitHub. The finished link lands in Telegram when it is ready for review.

GitHubGitHub

Runs inside GitHub Actions and opens the PR in your repo.

TelegramTelegram

Delivers the PR link straight to your chat.

Next upJiraLinearGitHub ProjectsGitLab IssuesBitbucket

FAQ

Common questions

No. Everything runs on GitHub Actions inside your own repository. Zero servers.

Final call

Start with one command.

Run the setup, answer a few prompts, and push when you're ready.

npx @elmiristic/agent-ready init