{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreif5tkx6wskd3lwb3x5mim4um2lnpx2r2f25c7hn2mtliwy3x5y2mu",
    "uri": "at://did:plc:pgryn3ephfd2xgft23qokfzt/app.bsky.feed.post/3mpq7iib4h6b2"
  },
  "path": "/t/it-s-the-architecture-stupid-why-prompt-engineering-won-t-fix-agents/175246?page=2#post_32",
  "publishedAt": "2026-07-03T08:43:41.000Z",
  "site": "https://discuss.huggingface.co",
  "tags": [
    "Zenodo",
    "Beyond Prompted Caution and Guardrails: Runtime-Enforced Pre-Action Cognition..."
  ],
  "textContent": "Another use case of ORCA, Action preflight\n\nZenodo\n\n### Beyond Prompted Caution and Guardrails: Runtime-Enforced Pre-Action Cognition...\n\nLLM agents are beginning to act, not merely answer. They call tools, write files, send messages,access private data, trigger workflows, and delegate work. Yet the reasoning that should precedeconsequential action often remains optional: hidden inside...",
  "title": "“It’s the Architecture, Stupid” — Why Prompt Engineering Won’t Fix Agents"
}