nobrainer-tech

AI dev & SDET. Practical AI, automation and software quality. Make Tech a No-Brainer. #NoBrainerTechFlow #LangflowMCP

[nobrainer-tech @nobrainer_tech]

50+ AI coding tools you could piece together, or one workflow for the agent you already use.

nobrainer-tech-flow carries a task through and shows what was delivered, what was checked and what is left to decide.

Free, open source: https://t.co/cmlNEWhBgb

#NoBrainerTechFlow

[nobrainer-tech @nobrainer_tech]

@0xKubi Ponytail jest dobry, ale generalnie da się wycisnąć z niego więcej. Tu moja wariacja nt. minimalistycznego zestawu workflow, który sprawia, że rzadziej wyzywam LLMy od najgorszych i nie trzeba pisać „make no mistakes” sprawdź sobie, może Ci podejdzie: https://t.co/9MFUPh2xHZ

[nobrainer-tech @nobrainer_tech]

@progXprog najlepiej ogarnąć pod siebie, mi się sprawdza to: https://t.co/cmlNEWhBgb ale tak jak pisali przedmówcy, im mniej tym lepiej :)

[nobrainer-tech @nobrainer_tech]

Langflow MCP 4.12.2 is live on npm.
This release was delivered end to end with nobrainer-tech-flow: API fixes, tests, review and publication.
Connect your AI agent to Langflow:
https://t.co/kmvbiMip5U
Workflow: https://t.co/97yVfV1bBz
#NoBrainerTechFlow

[nobrainer-tech @nobrainer_tech]

@imryven Would you keep the advisor and final verifier as separate roles? Your article argues for clean review context, while /advisor reads the whole session. I'd try task + diff + raw results first, then reveal the worker's explanation and compare the findings.

[nobrainer-tech @nobrainer_tech]

@karpathy The words I'd protect during simplification are "if", "unless" and "not yet verified". Drop those and a clearer sentence can become a stronger claim. Same for diagrams: I'd mark an unknown dependency rather than let an absent arrow imply there isn't one.

[nobrainer-tech @nobrainer_tech]

Before accepting an agent's change, check which acceptance IDs ran. A local dev check passed eight tests while five ID checks weren't selected. Match executed IDs to your criteria. Free nobrainer-tech-flow: https://t.co/97yVfV1bBz

[nobrainer-tech @nobrainer_tech]

An idea shouldn't stop at a plan.

nobrainer-tech-flow 2.1.0: idea → SDD → execution → checked result.

Plus plain technical English inspired by ASD-STE100, following @karpathy's suggestion.

Try one real task: https://t.co/97yVfV1bBz
#NoBrainerTechFlow

[nobrainer-tech @nobrainer_tech]

Before accepting an AI test fix, inspect the diff for skipped tests or weaker assertions. Rerun the failing case.

Our free nobrainer-tech-flow includes a review gate for the actual change:
https://t.co/97yVfV1bBz
#NoBrainerTechFlow

[nobrainer-tech @nobrainer_tech]

The page also gives you one instruction for https://t.co/q3sDDcrmqQ Flow: scope, split, delegate, verify. Paste it into your app's chat and the agent saves it for every new session.

Flow is free and open source: https://t.co/97yVfV1bBz

[nobrainer-tech @nobrainer_tech]

ChatGPT + Claude in Codex:
- high-stakes: GPT-6 Astra xhigh
- everyday: GPT-6.1 Sol xhigh

Sol gets 97% of Astra's orchestration score at 17% of the cost per task (list API prices).

Tick your plans and app to get your own picks. Free: https://t.co/of9uyO1qgP

[nobrainer-tech @nobrainer_tech]

@shb7el3alm My favourite response is „you are right” lol

[nobrainer-tech @nobrainer_tech]

nobrainer-tech-flow 2.0.2 is out: one prompt installs and updates it in any coding agent. No Python, no version to copy. It previews first, applies after you agree, and gives you the undo command.

Try it on a real task: https://t.co/97yVfV0DM1 #NoBrainerTechFlow

[nobrainer-tech @nobrainer_tech]

@marek_2k22 https://t.co/ZQZidkCjM8 sam sie na tą platformę przeniosłem, bo gumroad i naffy dawały za mało możliwości dostosowania pod własne potrzeby

[nobrainer-tech @nobrainer_tech]

Your next AI task shouldn't start with another pile of skills.

nobrainer-tech-flow 2.0 brings setup, context, planning and checks into one workflow. Load what the task needs; keep your main model.

Try it: https://t.co/97yVfV0DM1
#NoBrainerTechFlow

[nobrainer-tech @nobrainer_tech]

Don't start your next big Codex task with unused capacity.
I made my setup public: check real model context, see if MAX is available, configure 15 Luna slots, and let nobrainer-tech-flow keep MAIN in charge.
Try it first: https://t.co/F2zKoxhU9S #NoBrainerTechFlow

[nobrainer-tech @nobrainer_tech]

An agentic SDLC is not just more agents. It needs roles, adversarial review, fact checking and an integration owner. Jev and Laya may add judgment, but ownership stays with workflow evidence. What's the gate?
https://t.co/H1RJpNjOBo

[nobrainer-tech @nobrainer_tech]

Cloudflare's security-audit skill uses independent verification and machine-readable findings. That's a domain skill, not a full workflow. Jev and Laya may aid judgment; the acceptance contract stays explicit.
https://t.co/9yNkIz9VJn

[nobrainer-tech @nobrainer_tech]

An operating layer should make a workflow easier to trust: planning, routing, memory and evidence gates in one path. Jev and Laya can support a decision, but the core workflow should still stand without them. Where should a runtime stop?
https://t.co/OYAQEEltXp

[Tibo @thsottiaux]

GPT-6 Sol and Luna are out. Not only are they a very significant improvement across the board, but also in writing and general "you know when you try it" quality.

We are also permanently reducing the API price by 50% making both of them viable for a ton of new usecases and making your usage go further too, even on the subscriptions.

And one more thing. We are loading a banked reset into all accounts of our Plus, Pro and Business users. Let's go!

https://t.co/00DRh1sRrO