{"slug": "declarative-ai-agents-spec-once-generate-everywhere-with-acel", "title": "Declarative AI Agents: Spec Once, Generate Everywhere with ACEL", "summary": "A developer introduced ACEL (Agent Capability Expression Language), a declarative language for defining AI agents, built on the Rectified Pentachoron Framework (RPF). The language models agents as five aspects plus an autonomy modal, and includes a generator tool, acel-generator, that produces artifacts like AGENT.md and MCP configuration. The projects are MIT-licensed and aim to make agents governable, auditable, and portable.", "body_md": "Most agent frameworks still force you to write imperative code: wire tools, manage memory, hard-code oversight logic, and re-implement the same patterns for every new runtime. The result is brittle agents that are hard to review, version, audit, or port.\n\nA cleaner path is emerging: treat the agent as a **declarative contract**.\n\n**ACEL (Agent Capability Expression Language)** is a small, focused language for defining AI agents. It is built on the Rectified Pentachoron Framework (RPF) and models an agent as exactly five aspects plus a global autonomy modal:\n\n| Aspect | Role | Neglect pathology |\n|---|---|---|\ntelos |\nThe objective pursued | Aimless reactivity |\nworld_model |\nPresent apprehension of environment state | Ungrounded action (hallucination) |\nmemory |\nState retained across time | Amnesia |\ndeliberation |\nMapping situation → action | Thrashing / myopia |\nactuation |\nEffecting change on the environment | Paralysis |\n\nAutonomy is *not* a sixth aspect. It is a modal that qualifies every aspect (supervised perception, supervised deliberation, etc.). Oversight rules sit alongside it as first-class triggers (uncertainty thresholds, irreversible actions, cost limits, etc.).\n\nA minimal example looks like this:\n\n```\nagent research_assistant {\n  autonomy: SUPERVISED\n\n  telos {\n    goal: ACHIEVE \"comprehensive_report\" AND MAINTAIN \"factual_accuracy\"\n    priority: HIGH\n  }\n\n  world_model {\n    grounding: REQUIRED\n    percept web_search { type: RETRIEVE protocol: MCP freshness: 1h }\n  }\n\n  memory {\n    type: HYBRID\n    store short_term { capacity: 10000 ttl: 1h retrieval: RECENCY }\n    store long_term  { capacity: 1000000 retrieval: RELEVANCE }\n  }\n\n  deliberation {\n    mode: HYBRID\n    plan   { strategy: HIERARCHICAL depth: 4 replan_on_failure: true }\n    decide { strategy: MONTE_CARLO_TREE_SEARCH risk: 0.4 explore: 0.2 }\n    reflect { trigger: ON_ERROR over: long_term }\n  }\n\n  actuation {\n    effect document_store { type: MODIFY protocol: REST permissions: READ, WRITE }\n  }\n\n  oversight high_stakes {\n    trigger: UNCERTAINTY_ABOVE 0.3\n    action: ASK_APPROVAL\n    escalate_to: lead\n  }\n}\n```\n\nThe language enforces clean separations (percepts are read-only; effects are write; reflection must bind to a declared memory store) and produces a conformance report against the five-aspect basis. The `.acel`\n\nfile becomes the single source of truth — versionable, reviewable, and enforceable in CI.\n\nWriting the contract is only half the story. **acel-generator** (OpenAPI-Generator-style tooling for ACEL) turns that contract into concrete artifacts.\n\nPipeline:\n\n```\n.acel  →  parse + validate + RPF conformance  →  IR  →  generators  →  artifacts\n```\n\nCurrent built-in generators include:\n\n`AGENT.md`\n\n) + Mermaid architecture diagram`.well-known/agent-card.json`\n\n) for discovery and delegation`mcp.json`\n\n) with percepts as read tools and effects as write tools (plus `destructiveHint`\n\nannotations)Governance details that current interoperability protocols do not express (autonomy level, oversight, grounding, retention shape) are emitted under namespaced extensions so the gap stays visible rather than being papered over.\n\nDesign principles of the generator:\n\nACEL is deliberately *not* a runtime. It is the contract that binds the layers you already run (agent loop, control plane, gateway, tools/MCP, memory). Author-time validation shifts left; runtime enforcement compiles autonomy and oversight into gateway rules.\n\nBoth are MIT-licensed (with a narrow patent non-assertion on the language itself). The projects are early (alpha), but the conceptual clarity is already strong.\n\nIf you are tired of re-implementing the same agent skeleton for every framework, try writing the contract once and generating the rest. Declarative agents are not just cleaner code — they are a practical step toward agents that are governable, auditable, and portable by design.", "url": "https://wpnews.pro/news/declarative-ai-agents-spec-once-generate-everywhere-with-acel", "canonical_source": "https://dev.to/tposchel/declarative-ai-agents-spec-once-generate-everywhere-with-acel-88k", "published_at": "2026-08-10 13:32:00+00:00", "updated_at": "2026-08-10 13:49:34.028938+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-agents", "developer-tools", "ai-infrastructure"], "entities": ["ACEL", "Rectified Pentachoron Framework", "acel-generator", "MCP"], "alternates": {"html": "https://wpnews.pro/news/declarative-ai-agents-spec-once-generate-everywhere-with-acel", "markdown": "https://wpnews.pro/news/declarative-ai-agents-spec-once-generate-everywhere-with-acel.md", "text": "https://wpnews.pro/news/declarative-ai-agents-spec-once-generate-everywhere-with-acel.txt", "jsonld": "https://wpnews.pro/news/declarative-ai-agents-spec-once-generate-everywhere-with-acel.jsonld"}}