Components of a Coding Agent(magazine.sebastianraschka.com) |
Components of a Coding Agent(magazine.sebastianraschka.com) |
I think spec-driven generation is the antithesis of chat-style coding for this reason. With tools like Claude Code, you are the one tracking what was already built, what interfaces exist, and why something was generated a certain way.
I built Ossature[1] around the opposite model. You write specs describing behavior, it audits them for gaps and contradictions before any code is written, then produces a build plan toml where each task declares exactly which spec sections and upstream files it needs. The LLM never sees more than that, and there is no accumulated conversation history to drift from. Every prompt and response is saved to disk, so traceability is built in rather than something you reconstruct by scrolling back through a chat. I used it over the last couple of days to build a CHIP-8 emulator entirely from specs[2]. I have some more example projects on GitHub[3]
1: https://github.com/ossature/ossature
Now the coding agent starts fresh each time and its up to you to understand what you asked it and provide the feedback loop.
Instead of chat -> code, I think chat -> spec and then spec -> code is much more the future.
the spec -> code phase should be independent from any human. If the spec is unclear, ask the human to clarify the spec, then use the spec to generate the code.
What happens today is that something is unclear and there is a loop where the agent starts to uncover some broader understanding, but then it is lost the next chat. And then the Human also doesn't learn why their request was unclear. "Memories" and Agents files are all ducktape to this problem.
You talk with agent A it only modifies this spec, you still chat and can say "make it prettier" but that agent only modifies the spec, the spec could also separate "explicit" from "inferred".
And of course agent B which builds only sees the spec.
User actually can care about diffs generated by agent A again, because nobody wants to verify diffs on agents generated code full of repetition and created by search and replace. I believe if somebody implements this right it will be the way things are done.
And of course with better models spec can be used to actually meaningfully improve the product.
Long story short what industry misses currently and what you seem to be understanding is that intent is sacred. It should be always stored, preferably verbatim and always with relevant context ("yes exactly" is obviously not enough). Current generation of LLMs can already handle all that. It would mean like 2-3x cost but seem so much worth it (and the cost on the long run could likely go below 1x given typical workflows and repetitions)
I agree a dedicated layer for intent capture makes a lot of sense. I thought about that as well, I am just not fully convinced it has to be conversational (or free-form conversational). Writing a prompt to get the right spec change is still a skill in itself, and it feels like it'd just be shifting the problem upstream rather than actually solving it. A structured editing experience over specs feels like it'd be more tractable to me. But the explicit vs inferred distinction you mention is interesting and worth thinking through more.
I'm using something similar-ish that I build for myself (much smaller, less interesting, not yet published and with prettier syntax). Something like:
a->b # b must always be true if a is true
a<->b # works both ways
a=>b # when a happens, b must happen
a->fail, a=> fail # a can never be true / can never happen
a # a is always true
Product.alcoholic? Product in Order.lineItems -> Order.customer.can_buy_alcohol?
u1 = User(), u2=User(), u1 in u2.friends -> u2 in u1.friends
new Source() => new Subscription(user=Source.owner, source=Source)
Source.subscriptions.count>0 # delete otherwise
How does the human intervention work out? Do you use a mix of spec and audit editing to get into the ready to generate state? How high is the success/error rate if you generate from tasks to code, do LLMs forget/mess up things or does it feel better?
The spec driven approach is potentially better for writing things from scratch, do you have any plans for existing code?
> How does the human intervention work out? Do you use a mix of spec and audit editing to get into the ready to generate state?
Yes, the flow is: you write specs then you validate them with `ossature validate` which parses them and checks they are structurally sound (no LLM involved), then you run `ossature audit` which flags gaps or contradictions in the content as INFO, WARNING, or ERROR level findings. The audit has its own fixer loop that auto-resolves ERROR level findings, but you can also run it interactively, manually fix things yourself, address the INFO and WARNING findings as you see fit, and rerun until you are happy. From that it produces a toml build plan that you can read and edit directly before anything is generated. You can reorder tasks, add notes for the LLM, adjust verification commands, or skip steps entirely. So when you run `ossature build` to generate, the structure is already something you have signed off on. There's a bit more details under the hood, I wrote more in an intro post[1] about Ossature, might be useful.
> The spec driven approach is potentially better for writing things from scratch, do you have any plans for existing code?
Right now it is best for greenfield, as you said. I have been thinking about a workflow where you generate specs from existing code and then let Ossature work from those, but I am honestly not sure that is the right model either. The harder case is when engineers want to touch both the code and the specs, and keeping those in sync through that back and forth is something I want to support but have not figured out a clean answer for yet. It's on the list, if you have any thoughts please feel free to open an issue! I want to get through some of the issues I am seeing with just spec editing workflow (and re-audit/re-planning) first, specifically around how changes cascade through dependent tasks.
Regarding success rate, each task requires a verification command to run and pass after generation and if it fails, a separate fixer agent tries to repair it using the error output. The number of retry attempts is configurable. I did notice that the more concise and clear the spec is the more likely it is for capable models to generate code that works (obviously) but that's what auditing is supposed to help with. One interesting case about the chip-8 emulator I mentioned above is that even mentioning the correct name of the solution to a specific problem was not enough, I had to spell out the concrete algorithm in the spec (wrote more details here[2]). But the full prompt and response for every task is saved to disk, so when something does go wrong one can read the exact prompt/response and fix-attempts prompt/response for each task.
But the idea is similar in that I start with a spec and feed the LLM context that is a projection of the code and spec, rather than a conversation. The context is specific to the specific workflow stage (eg planning needs different context to implementing) and it doesn’t accumulate and grow (at least, the growth is limited and based on the tool call loop, not on the entire process).
My main goals are more focused context, no drift due to accumulated context, and code-driven workflows (the LLM doesn’t control the RPI workflow, my code does).
It’s built as a workflow engine so that it’s easy for me to experiment with and iterate on ideas.
I like your idea of using TOML as the artifact that flow between workflow stages, I will see if that’s something that might be useful for me too!
Any reason you’ve opted for custom markdown formats with the @ syntax rather than using something like frontmatter?
Very conscious that this would prevent any markdown rendering in github etc.
> Yeah, I did briefly consider front-matter, but ended up with inline @ tags because I thought it kept the entire document feeling like one coherent spec instead of header-data + body, front matter felt like config to me, but this is 0.0.1 so things might change :)
I notice you support ollama. Have you found it effective with any local models? Gemma 4?
I'm definitely going to play with this.
Well, for the first problem, if an AI can generate the code in a day or a week, the world hasn't moved very much in that time. (In the future, if everything is moving at the speed of AI, that may no longer be true. For now it is.)
The second problem... if Ossature (or equivalent) warns you of gaps rather than just making stuff up, you could wind up with iterative development of the spec, with the backend code generation being the equivalent of a compiler pass. But at that point, I'm not sure it's fair to call it "waterfall". It's iterative development of the spec, but the spec is all there is - it's the "source code".
It's also a good example that you can turn any useful code component that requires 1k LOC into a mess of 500k LOC.
Since my agent works over email, the core agent loop only processes one message then hits the send_reply tool to craft a response. Then the next incoming email starts the loop again from scratch, only injecting the actual replies sent between user and agent. This naturally prunes the context preventing the long context window problem.
I also had a challenge deciding what context needs injecting into the initial prompt vs what to put into tools. Its a tradeoff between context bloat and cost of tool lookups which can get expensive paying per token. Theres also caching to consider here.
Full writeup is here if anyone is interested: https://www.healthsharetech.com/blog/building-alice-an-empow...
People have been doing that for over a year already? GLM officially recommends plugging into Claude Code https://docs.z.ai/devpack/tool/claude and any model can be plugged into Codex CLI (it's open source and can be set via config file).
Unless I'm misunderstanding what's being described here, running Claude Code with different backend models is pretty common.
https://docs.z.ai/scenario-example/develop-tools/claude
It doesn't perform on par with Anthropic's models in my experience.
If you want to play with the basic building blocks of coding agents, check out https://github.com/OpenHands/software-agent-sdk
https://github.com/shareAI-lab/learn-claude-code/tree/main/a...
I found it excellent in explaining a CC-like coding agent in layers.
I suspect that more could be done in terms of translating semi-naive user requests into the steps that a senior developer would take to enact them, maybe including the tools needed to do so.
It's interesting that the author believes that the best open source models may already be good enough to complete with the best closed source ones with an optimized agent and maybe a bit of fine tuning. I guess the bar isn't really being able to match the SOTA model, but being close to competent human level - it's a fixed bar, not a moving one. Adding more developer expertise by having the agent translate/augment the users request/intent into execution steps would certainly seem to have potential to lower the bar of what the model needs to be capable of one-shotting from the raw prompt.
For a preview of what it'd be like, just tell your AI chat app that you'll run bash commands for it, and please change the app in your "current directory" to "sort the output before printing it", or some such request.
Why do you think that is the case? Is Anthropic's models just better or do they train the models to somehow work better with the harness?
If you want to look at some of the tooling and process for this, check out verifiers (https://github.com/PrimeIntellect-ai/verifiers), hermes (https://github.com/nousresearch/hermes-agent) and accompanying trace datasets (https://huggingface.co/datasets/kai-os/carnice-glm5-hermes-t...), and other open source tools and harnesses.
I also find OpenCode to be drastically better than Claude Code, to the extent that I'm buying OpenRouter API credits rather than Claude Max because Claude Code just isn't good enough.
I'm frankly amazed at what OpenCode can do with a few custom commands (just for common things like doing a quality review, etc.), and maybe an extra "agent" definition or two. For many projects even most of this isn't necessary. Often I just ask it to write an AGENTS.md that encapsulates a good development workflow, git branch/commit policy, testing and quality standards, and ROADMAP.md plus per milestone markdown files with phases and task tracking, and this is enough.
I'm somewhat interested in these more involved harnesses that automated or enforce more, but I don't know that they'd give me much that I don't have and I think they'd be tough to keep up with the state of the art compared to something less specific.
Ossature uses two markdown formats, SMD[1] for describing behavior and AMD for structure (components, file paths, data models). AMDs[2] link back to their parent SMD so behavior and structure stay connected. Both are meant to be written, reviewed, and/or owned by humans, the LLM only reads the relevant parts during generation. One thing I am thinking about for the future is making the template structure for this customizable per project, because "spec" means different things to different teams/projects. Right now the format is fixed, but I am thinking about a schema-based way to declare which sections are required, their order, and basic content constraints, so teams can adapt the spec structure to how they think about software without having to learn a grammar language to do it (though maybe peg-based underneath anyway, not sure).
The formal approach you describe is probably more precise for expressing system properties. Would be interesting to see how practical it is to maintain it as a project grows.
Do you have a source? Claude Code is the only genetic system that seems to really work well enough to be useful, and it’s equipped with an absolutely absurd amount of testing and redundancy to make it useful.
I agree that, this is what makes it not waterfall. You're iterating on the spec and not backtracking from broken code. The spec is the "source code", replanning and rebuilding is just "recompiling".
So, yes, it can work.
Context management, both within and across sessions, seems the bigger issue. Without the agent supporting this, you are at the mercy of the model compacting/purging the context as needed, in some generic fashion, as well as being smart enough to decide to create notes for itself tracking what it is doing, etc.
Apparently CC is 512K LOC, which seems massively bloated, but I do think that things like tools, skills, context management and subagents are all needed to effectively manage context and avoid the issues that might be anticipated by just telling the model it's got a bash tool, and go figure.
Okay sure it’s technically more than just bash, but my own for-fun coding agent and pi-coding-agent work this way. The latter is quite useful. You can get surprisingly far with it.
It's just that we're lazy. After being able to chat, I don't see people going back. You can't just paste some error into the specs, you can't paste it image and say it make it look more like this. Plus however well designed the spec, something like "actually make it always wait for the user feedback" can trigger changes in many places (even for the sake of removing contradictions).
1. You can write a spec that builds something that is not what you actually wanted
2. You can write spec that is incoherent with itself or with the external world
3. You can write a spec that doesn't have sufficient mechanical sympathy with the tooling you have and so it requires you to all spec out more and more of the surrounding tech than you practically can.
All of those issues can be addressed by iterating on the spec with the help of agents. It's just an engineering practice, one that we have to become better at understanding
The third point is harder. You still need to know your tooling well enough to write a spec that works with it. That part hasn't gone away.
Say "I want HN client for mobile", "must notify me about comments", you see it and you add "should support dark mode". Can you see how that is much less than anything in any programming language?
I also make individual tasks md files (task.md) which makes them capable of carrying intent, plan, but not just checkbox driven "- [ ]" gates, they get annotated with outcomes, and become a workbook after execution. The same task.md is seen twice by judge agents which run without extra context, the plan judge and the implementation judge.
I ran tests to see which component of my harness contributes the most and it came out that it is the judges. Apparently claude code can solve a task with or without a task file just as well, but the existence of this task file makes plans and work more auditable, and not just for bugs, but for intent follow.
Coming back to user intent, I have a post user message hook that writes user messages to a project scoped chat_log.md file, which means all user messages are preserved (user text << agent text, it is efficient), when we start a new task the chat log is checked to see if intent was properly captured. I also use it to recover context across sessions and remember what we did last.
Once every 10-20 tasks I run a retrospective task that inspects all task.md files since last retro and judges how the harness performs and project goes. This can detect things not apparent in task level work, for example when using multiple tasks to implement a more complex feature, or when a subsystem is touched by multiple tasks. I think reflection is the one place where the harness itself and how we use it can be refined.
claude plugin marketplace add horiacristescu/claude-playbook-plugin
source at https://github.com/horiacristescu/claude-playbook-plugin/tree/mainThe judge finding is interesting though. Right now verification during build for each task in Ossature is command-based, compile, tests, that kind of thing. A judge checking spec-to-code fidelity rather than (or maybe in addition to?) runtime correctness is worth thinking about.
One interesting data point - I counted word count in my chat messages vs final code and they came out about 1:1, but in reality a programmer would type 10x the final code during development. From a different perspective I found I created 10x more projects since I relied on Claude and my harness than before. So it looks user intent is 10x more effective than manual coding now.
I just asked Claude, and apparently CC makes it's bash tool available on all platforms it runs on (Linux, macOS, Windows WSL, Git for Windows), and doesn't do platform-specifc filtering of bash commands, which would seem to make for some interesting incompatibilities - GNU utils (sed, grep, find) on Linux and Windows, but BSD variants on macOS.