Pierwszy działający efekt pokazuję zwykle w 2-4 tygodnie. Zamiast slajdów i obietnic dostajesz rozwiązanie, które od razu możesz przetestować u siebie w firmie.

Ile kosztuje współpraca?

Szybki start produktu to od 15 000 PLN jednorazowo. Później dostępna jest opieka na abonament (od 5 000 PLN/mies) na rozwój i utrzymanie, oraz pakiet automatyzacji (od 3 000 PLN/mies) na kolejne usprawnienia. Dokładną wycenę podaję po 30-minutowej rozmowie.

Nie znam się na technologii, dam radę?

Tak. Rozmawiamy po ludzku, bez żargonu. Ty mówisz, co Cię uwiera w firmie, a ja zajmuję się całą techniczną stroną i dostarczam gotowe do użycia narzędzie.

Dla kogo jest ta oferta?

Dla firm, które tracą godziny na powtarzalną, ręczną pracę (wyceny, raporty, odpowiedzi, obsługa zamówień) i chcą to usprawnić bez budowania własnego zespołu programistów.

Jak zacząć współpracę?

Umów bezpłatną rozmowę na https://bartoszgaca.pl/rozmowa/. Opowiadasz, co chcesz usprawnić, a Ty dostajesz konkretny plan: co da się zrobić, ile to zajmie i ile kosztuje.

Claude Code SDK and GitHub Action: Automation Without an Interface

TL;DR: The Claude Code SDK is programmatic access to the Claude agent in headless mode — no GUI, no clicks, just bash and pipes. Built on the same engine as the CLI, it lets you integrate AI into CI/CD pipelines, scheduled jobs and webhooks. The GitHub Action adds Claude as a reviewer to your pull requests.

TL;DR

The Claude Code SDK is programmatic access to the Claude agent in headless mode: no GUI, no clicking, just bash and pipes. It is built on the same philosophy as Unix tools: plug it in anywhere, chain it, build CI with it. The Claude GitHub Action (open source) runs on top of the SDK. Comment on a GitHub issue and Claude implements the feature or reviews the PR on its own.

What headless mode is

The Claude Code SDK is not another wrapper around a chat API. It is access to the agent, the same one that runs in the terminal, in headless mode.

"It is a way to programmatically access the power of the Claude Code Agent in headless mode. This is a new kind of primitive and a new kind of building block that allows you to build applications that just weren't possible before."

The difference is fundamental: the agent has the project context, understands the file structure and can write code, not just generate text.

The Unix philosophy: why it works

The SDK is designed like a Unix tool. That means:

You can plug it in anywhere bash or a terminal runs
Data comes in through a pipe, data goes out through a pipe
You can build complex processing chains

This is not a metaphor. The commands below work literally.

SDK commands, straight from the talk

Basic call


claude -p "write me a fibonacci sequence generator" --allowedTools Write

The --allowedTools flag (also spelled --allowed-tools) lists the tools that run without a permission prompt, for example writing files on the local system. Without it, Claude asks before every write.

Log analysis through a pipe


cat app.log | claude -p

Instead of reading logs by hand, pipe them to Claude and it summarizes the errors. Works on any text output.

Parsing system output


ifconfig | claude -p

The same pattern works for ifconfig, df, ps aux, anything whose output is hard to interpret by hand.

JSON output: the key to automation


claude -p "task" --output-format json

Instead of plain text, Claude returns structured JSON. This is what makes the SDK useful programmatically: you can parse the output and build the next steps of the pipeline on top of it. The documented formats are text, json and stream-json.

CI/CD use cases (named explicitly)

Code review: Claude reviews code in the pipeline
Custom linters: for things that cannot be defined programmatically, Claude does it for you
Chatbots: powered by the Claude Code agent (possible, not demonstrated)
Remote environments: writing code in isolated environments (possible, not demonstrated)

GitHub Action: the live demo

The GitHub Action is a ready implementation built on top of the SDK. It is open source (anthropics/claude-code-action).

It can:

Create new features based on an issue
Review pull requests
Triage bugs

Workflow (exactly as presented)

Issue #1: power-ups for a quiz app.

Issue text: "add a powerup option in the config, 50/50 elimination for the skip question, it should award user points even though the question was skipped, users should be able to configure this from the config page"

The issue was deliberately short on detail, leaving a lot of creative room for the agent.

Issue #2: a per-question timer.

A global countdown existed, a per-question timer was missing.

What happened:

A comment on the issue: @claude please implement this feature
After about 4 to 5 seconds Claude replied with a confirmation comment and a link to the job run
The GitHub Action logs showed the JSON output from the SDK
Claude created its own TODO list and went through it item by item
Both issues were processed in parallel
The same mechanism works for PR review: Claude comments on the changes

Announced, not demonstrated

Python bindings for the SDK
TypeScript bindings for the SDK

How to implement it

Install Claude Code and run claude in the terminal
Test a pipe: echo "explain this" | claude -p
Test log analysis: cat your-log.txt | claude -p
Add --output-format json and parse the output in a script
Install the Claude GitHub Action on the repository (run /install-github-app inside Claude Code)
Create an issue describing a feature and comment @claude please implement this