Djordje Lukic

@alexdesousa @aheritier @Docker @dgageot On a une commande "docker agent new", ca lance un agent special qui sait creer des agents pour toi. Je l'utilise souvent

@aheritier @Docker @dgageot @rumpl Créer des agents IA, c'est déjà complexe. Je me demande si on pourrait tirer parti de modèles pré, entraînés pour accélérer ce processus. Ça pourrait vraiment changer la donne, non ?

@docker Agent - comment simplifier encore plus la création d'agents IA ? par @dgageot and @rumpl youtube.com/watch?v=WI_aNY…

@MrAhmadAwais @CommandCodeAI We've seen similar things in docker-agent with haiku (can't remember which version) so we added this github.com/docker/docker-…

how did we make deepseek outperform opus 4.7? i've been thinking about why "open model bad at tool calling" is almost always a harness problem, not a model problem. context: spent the two days looking at billions of tokens in @CommandCodeAI (tb open source ai cli) using deepseek. I ended up writing a tool-input repair layer. the trigger was watching deepseek-flash fail on the simplest /review run, every shellCommand and readFile call bouncing back with a raw zod issues blob, the model unable to recover because the error wasn't in a form it could read. by the end deepseek v4 pro was beating opus 4.7 6/10 times on our internal evals. a few things i learned that feel general: 1/ the failure modes aren't random they're a small finite compositional set. across deepseek-flash, deepseek v4 pro, glm, qwen, the same four mistakes repeat almost exactly: - sending `null` for an optional field instead of omitting it - emitting `["a","b"]` as a json *string* instead of an actual array - wrapping a single arg in `{}` where the schema expected an array (an "empty placeholder") - passing a bare string where an array was expected (`"foo"` instead of `["foo"]`) four repairs, ~30-100 lines each, ordered carefully (json-array-parse must run before bare-string-wrap or `'["a","b"]'` becomes `['["a","b"]']`). that is the whole catalogue. when i hear "this open source model can't do tool calls" i now assume one of those four, and so far that's been right ~90% of the time. 2/ the funniest failure mode is also the most revealing. deepseek-flash, when asked to edit or write a file, sometimes emits the path as a *markdown auto-link*: filePath: "/Users/x/proj/[notes.md](http://notes. md)" our writeFile tool obediently trued creating files literally named `[notes.md](http://notes .md)` until we caught it. this is not a hallucination. it's the post-training chat distribution leaking through the tool boundary the model has been rewarded for auto-linking in conversational output, and is applying that prior in a context where it makes no sense. the fix is two regex lines that unwrap only the degenerate case where link text equals url-without-protocol real markdown like `[click](https://x .com)` passes through untouched. this is also conditioning of their own tools during RL which were different from all other tools we write and ofc can't predict. "tool confusion" is a more useful frame than "capability gap." the model knows how to format a path. it just hasn't been told clearly enough that this path is going to fopen, not into a chat bubble. so we encode that hint at the schema level `pathString()` instead of `z.string()` and the leak is plugged for every path field at once. 3/ the design choice that mattered was inverting preprocess-then-validate to validate-then-repair. my first attempt was the obvious one: a preprocessing pass that normalized inputs (strip nulls, parse stringified arrays, etc.) before zod ever saw them. it broke immediately, writeFile content that *happened* to be json-shaped got rewritten before it hit disk. silent corruption, easy to miss in a smoke test. then i made it less greedy - parse the input as-is. if it succeeds, ship it. valid inputs are never touched. - on failure, walk the validator's own issue list. for each issue path, try the four repairs in order until one applies. - parse again. on success, log `tool_input_repaired:${toolName}`. on failure, log `tool_input_invalid:${toolName}` and return a model-readable retry message. the structural insight here is: when you preprocess, you encode a prior about what's broken. when you let the validator complain first, the schema is the prior, and you only spend repair budget at the exact paths the schema actually disagreed at. the validator is doing the work of localizing the bug for you. it's the same shape as cheap-then-careful everywhere else try the fast path, fall back on evidence. (this also gives you per-tool telemetry for free. you can watch repair rates per (model, tool) and notice when a model regresses on a specific contract before users do.) 4/ shape invariants and relational invariants need different fixes. the four repairs above all handle shape problems wrong type, missing key, wrong container. but read_file had a *relational* invariant: "if you provide offset, you must also provide limit, and vice versa." deepseek kept calling `readFile({ absolutePath, limit: 30 })` and getting an `ERROR:` back. you can't fix this with input repair, because each field is independently valid the bug is in the relationship between them. so i taught the function the model's intent instead. `limit` alone → `offset = 0`. `offset` alone → `limit = 2000` (matches common read tool ops default). then surfaced the decision back to the model in the result: "Note: limit was not provided; defaulted to 2000 lines. To read more or fewer lines, retry with both offset and limit." no `Error:` prefix, so the tui doesn't paint it red. the model sees what we picked and can self-correct on the next turn if our guess was wrong. transparency over silent magic wins big. repair where you can. extend semantics where you can't. surface the choice either way. zoom out: a lot of what looks like model capability is actually contract design. a strict schema is a choice with a cost it filters out noise, but it also filters out recoverable noise from any model that hasn't memorized the exact json contract you happened to pick. the largest commercial models eat that cost invisibly and are linient on tool calling because they've seen enough of every contract during pretraining; open models pay it loudly and get dismissed for it. the harness is where you mediate between distributions. four small repairs (i'm sure more to follow as we have three more merging today), two regex lines for auto-links, one relational default, one prefix change. the model didn't change. the contract got more forgiving in exactly the places it needed to be. deepseek v4 pro now beats opus 4.7 6/10 times on our internal evals. imo "skill issue" applies to the harness more often than the model.

What I learned from patching Docker Engine default seccomp profile for CVE-2026-31431 (Copy Fail) 1. If a seccomp rule already filters an argument (like AF_VSOCK), it's just a matter of adding a second negation for the AF_ALG, right? Wrong! These are two rules that ...

@glenntoews @cursor_ai @opencode @claudeai @OpenAI I made it so I'm biased github.com/docker/docker-…

Who has the best agent harness right now, and why? @cursor_ai SDK, @opencode, @claudeai Agent SDK or @OpenAI Codex ? I think no one cares about Gemini CLI atp?

@RohitPaulK I'm curiois to know how good/bad is docker agent at acp github.com/docker/docker-…

Been working on this over the previous week! I've got a test harness ready, some basic checks in place, 3 agents wired up and a UI to display checks. Now to add a bunch more checks & make it easy to understand what each failure means

@Docker Agent 1.46.0 is out with full support of @AnthropicAI Opus 4.7 github.com/docker/docker-…

@grono_dev Rude for poles

Every time you open Neovim, you’re executing arbitrary code from random GitHub repos. Don’t pin commits? You’re running whatever some guy in Poland pushed while you were asleep. Pin them? Your config slowly rots. How are people not losing their minds over this?

@_so_bored Essaye docker agent, meme pas besoin de code, uniquement un fichier yaml github.com/docker/docker-…

Tout le monde utilise LangChain pour ses agents IA alors qu'un framework 10x plus simple existe Je parle d'Agno Le constat : - Un agent = une classe, quelques lignes - Multi-agent = natif, pas un bricolage - Le code se lit comme du pseudo-code Mais le vrai game changer c'est dans claude code Tu colles Agno, l'IA comprend du premier coup Pas de debug de 2h parce que le LLM a halluciné une abstraction LangChain dans Claude Code, le modèle se perd dans les 45 couches d'abstraction et invente des méthodes qui existent pas J'ai refait un agent de prospection dessus Même résultat. 4x moins de code

🎉 Excited to announce my new book! Operational AI with Docker With @manvar_harsh | Published by @PacktPublishing 🐳 docker model run 🔗 docker mcp gateway run 🤖 docker agent run 📦 docker sandbox run Launching May 11 🗓️ Pre-order 👇 amazon.com/dp/B0GTPYRT9Z #Docker

@grono_dev Joke aside, I want to try it

Starting to look good!

@aheritier @DevoxxFR Oh ya, so very ready

Are you ready for @DevoxxFR in 2 weeks ?!?!

@ibuildthecloud Jean-Claude

I like referring to my coding agent as Claude. I like it having a name and an identity. But I don't use Claude code. I need a name for my coding agent.

@fabienpenso I figured :)

@rumpl The httpd is so you can restart but not loose sessions

Moltis: 220k Rust LOC in 6 weeks, 2.2k⭐️ Arbor: 70k in 11 days, 350⭐️ Polyphony: 35k in 3 days, not even announced yet Built solo. Not toys, production-grade, tested, deterministic. AI is a generational shift bigger than the internet.

@fabienpenso No worries. Didn't know it was fixed

@rumpl New release fixes this. Sorry :(

@noperator You can keep the name :) github.com/docker/cagent/…

@rumpl yikes. nasty name collision :(

cagent: An agent sandbox that allows Docker-in-Docker. I use this for development and security testing. Work in progress but it's useful and ergonomic for my use cases. github.com/noperator/cage…

@noperator All good :) Internally we have an agent we call Nicolas Cagent :)

@noperator :) github.com/docker/cagent

My favorite feature of cagent: masking .git/ as read-only. Super easy for me to see exactly what the agent changed. In the worst case, I can always restore any code that the agent accidentally blasted away.

@badlogicgames A trick I do in cagent is to give the system multiple llms and they are randomly selected on each loop, works well when opus or gpt get stuck in a "but wait" loop #L50" target="_blank" rel="nofollow noopener">github.com/docker/cagent/…

Don't let people tell you that switching model mid session is a bad idea. It's super useful to quickly switch from say Opus to Codex to get a second opinion based on the exact same context without any ceremony like starting a new session, spawning a subagent, etc.