01 · the anvil
Anvil
issue → spec
Hammers a raw GitHub issue into an implementation-ready spec. Agents read the assignment against real code, hunting gaps and ambiguities — and a human approves every write.
- read-only analysis
- human-in-the-loop
~/jiri-studnicka $ whoami
I'm Jiří Studnička. I spent twenty years building backends for banks, insurers and energy companies. Now I build systems where AI agents write, test and ship production software — and I vouch that it holds.
// building agentic systems since GPT-3.5 — before it was cool
// depth −7 m · sediments of practice
Studnička is Czech for a little well. It took twenty years of enterprise strata to dig — and that is exactly the depth every agent I set loose on production code draws from.
IIS Tábor — Gupta and .NET. The craft begins.
Capgemini, Minerva, RADIUM. Co-owner and team leader at 4Leaders.
Scala, Akka, Play, Kafka, MongoDB. Functional programming, monads, event sourcing. A bank built from scratch.
Scala and Akka Persistence for the energy sector — event sourcing in live production.
Spark, Kafka, HBase, Hadoop, Hive. Integrating the primary systems and DWH of a multinational group.
Java, .NET, Python. Spring Boot, Oracle, Azure Synapse, Spark.
I understood the craft had just changed. I've been doing agentic engineering full-time ever since. You don't throw seniority away — you reforge it.
“Most AI engineers discovered Claude Code yesterday. Meanwhile I'm working out how to keep a fleet of agents running on a production repo for three days straight.”
// depth −18 m · the tool seam
I named my tools after the smithy. Not for the romance — because agentic engineering is a craft: it has a process, tools and quality control. This is my production cycle.
01 · the anvil
issue → spec
Hammers a raw GitHub issue into an implementation-ready spec. Agents read the assignment against real code, hunting gaps and ambiguities — and a human approves every write.
02 · the forge
issue → pull request
An orchestration platform: agents run through spec, implementation, code review, validation and CI loops — all the way to merge and the checks beyond it. No “trust us”: every requirement has its ledger.
03 · the quench
PR → proof
Self-healing testing in a real browser. When a selector drifts, the agent understands the original intent, the test completes — and the scenario repairs itself to run deterministically next time.
A spec-driven CLI that holds agents to their word: every requirement carries a stable REQ-ID and verifiable evidence in tests. Drift between spec and reality is measured, not guessed.
A library of 59 commands and 25 skills for Claude Code and Codex — paired audits and fixes for security, performance, a11y and more. Standardised agentic workflows.
Meeting recording → ready-made GitHub issues. Gemini reads the video including the screen, Claude writes issues in the repository's context, screenshots attach themselves.
…and more: agent-system (orchestration via Telegram), MCP servers, invoiceAI.
Spec-first. An agent without a proper assignment is a random code generator.
Proof, not vibes. Every requirement has a test, evidence and a screenshot.
Human-in-the-loop. AI writes. A senior vouches.
// depth −29 m · load-bearing masonry
software house · co-owner
An agentic software house: AI agents write the code, seniors own architecture, quality and security. Enterprise systems, AI integrations, legacy project rescues.
cognera.cze-commerce · co-owner
nopCommerce experts since 2010. Over 50 e-shops, over 100 custom plugins, B2B and B2C, ERP and CRM integrations.
nopshop.cz// depth −42 m · the waterline
Looking for someone who actually runs agentic development — not someone who just posts about it on LinkedIn? Lower the bucket.