Show HN: Kastor – Terraform-style specs for AI agents

(github.com)

23 points by weirdguy2 hours ago

6 comments

  • empath751 hour ago
    The problem right now with something like this is you're trying to nail jello to a wall. People haven't figured out what an agent is yet, and trying to crystalize what people happen to be doing right now means in a few months, you're going to be obsolete.
    • weirdguy1 hour ago
      I agree with you<p>don&#x27;t think the &quot;agent&quot; abstraction is stable enough to standardize the behavior&#x2F;runtime layer yet.<p>thing I&#x27;m trying to test is smaller: is that possible to standardize the outer contract around agents?<p>what inputs does it accept? what outputs does it promise? what model does it use? what prompt&#x2F;template does it depend on? what tools can it call? what target should it be built or applied to?<p>so here it is intentionally not trying to describe the full control loop or become a runtime. it&#x27;s more like a source-of-truth layer for the parts that are already showing up everywhere, even if the runtime patterns keep changing underneath.<p>I might be wrong, but my bet is that the runtime layer will keep changing, while the need for reviewable&#x2F;diffable agent contracts will not.
    • reactordev1 hour ago
      You mean Claude…
      • giancarlostoro18 minutes ago
        Claude is a model, an agent is more like the full solution. Some would call Claude Code itself an agent. Agents do things, models are just models until you use them.
  • handfuloflight1 hour ago
    The website is down?
    • weirdguy1 hour ago
      oh, it&#x27;s not, just root domain (non-www) without redirect
      • handfuloflight1 hour ago
        That&#x27;s not working either.
        • weirdguy1 hour ago
          try with www, should work, the apex zone still did not propagate for whatever reason
  • Leewen1 hour ago
    works for me. terraform for agents? neat
    • weirdguy1 hour ago
      thanks — that’s the direction. current version is still just validate + LangGraph codegen, but plan&#x2F;apply&#x2F;state is the part I’m trying to get to next
  • Dan-SC9 minutes ago
    [flagged]
  • rnagulapalle1 hour ago
    [flagged]
  • weirdguy2 hours ago
    hey HN, it&#x27;s first time posting something here.<p>I’m building Kastor: Go CLI&#x2F;declarative language&#x2F;config for AI agents.<p>the motivation: agent definitions often end up spread across framework code, prompt files, tool files&#x2F;mcps, platform UI settings, and env vars. that makes them hard to review, diff, reproduce or move.<p>Kastor is an attempt to put a source-of-truth layer above that.<p>right now the working proof of concept is narrow:<p>- .agent &#x2F; .tool &#x2F; .prompt files - HCL parser + validation - dependency&#x2F;reference checks - LangGraph codegen - runnable weather example - runnable content scheduler example<p>the long-term direction is Terraform-ish:<p>- build: compile to framework code - plan&#x2F;apply: reconcile hosted platform agents - state: track remote resources and detect drift<p>I’m deliberately not trying to build another agent runtime. the thing I’m trying to validate is whether agents need something closer to IaC: versionable, reviewable, declarative source of truth.<p>would appreciate any kind of feedback, especially on the language&#x2F;design.