agentleFS
joaoajmatosjoaoajmatos/ze/.agents/skills/system-one-models/SKILL.md

system-one-models

Skill2 starsChanged 47 days agoUnlicenseView on GitHub

Design AI-powered software with System One models (e.g. TypeSafe Jev): fast typed judgments (Choice, Score, Noul) with calibrated probabilities, not text generation. Covers design workflow, primitives, input structuring, LLM integration patterns, and empirical tradeoffs. Use when routing, ranking, extracting, verifying, guarding, or scoring unstructured input.

The file

We split this skill into the rules it gives an agent

11 rules. About 0 are common, 11 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈11

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    When working with TypeSafe's Jev, start with the documentation index to discover relevant pages and cookbooks.

    line 31
  2. Distinctive:

    When working with another System One provider, use their official docs.

    line 34
  3. Distinctive:

    Start from the behavior the user wants: what will the application show, select, change, or hand off?

    line 48
  4. Distinctive:

    When brainstorming or choosing an architecture, consider more than classification.

    line 53
  5. Distinctive:

    Verify and escalate: Check specific claims or fields against their evidence.

    line 74
  6. Distinctive:

    Ask one narrow, coherent judgment per question.

    line 103
  7. Distinctive:

    Keep the needed answers available.

    line 110
  8. Distinctive:

    Use probabilities and confidence to guide behavior, with thresholds evaluated on the user's data and consequences.

    line 123
  9. Distinctive:

    Keep policy explicit and raw judgments reusable.

    line 130
  10. Distinctive:

    Test representative cases and the resulting application behavior.

    line 141
  11. Distinctive:

    Treat cookbook thresholds and demo results as examples to evaluate, not universal rules or permanent model limitations.

    line 144

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Build with System One models

System One models make fast, structured judgments that software can consume directly. They return typed answers and probabilities rather than generated text or reasoning explanations. Code owns the workflow; the model supplies programmable common sense where ordinary code needs semantic understanding.

Before designing your first workflow, read references/empirical-findings.md. It captures the real-world tradeoffs, common failure modes, input structuring pitfalls, and model behavior quirks discovered during integration research. These apply broadly across System One models and save you debugging the same sharp edges from scratch.

Read the docs

The live docs for your System One model are the source of truth. This skill gives direction; the provider's docs carry current concepts, prompting guidance, API contracts, SDK usage, models, limits, and worked examples.

  • Providers define their own primitives, request/response shapes, SDKs, and model availability. Read their documentation before writing integration code.
  • When working with TypeSafe's Jev, start with the documentation index to discover relevant pages and cookbooks. Use targeted reads rather than loading the entire site.
  • When working with another System One provider, use their official docs.

Quick reference for TypeSafe (Jev)

TaskStart here; follow the relevant details
Understand the programming modelSystem One, building guide
Explore what to buildUse-case map, then relevant cookbooks from the index
Prepare inputs and questionsState, primitives, then the chosen primitive's page
Decide how to handle uncertaintyConfidence
Write API codeHTTP API, Python SDK, or JavaScript SDK

Find the useful shape

Start from the behavior the user wants: what will the application show, select, change, or hand off? Work backward to the judgments it needs. Keep known rules, calculations, exact lookups, and execution in code. Preserve the user's chosen stack and scope; add the System One model where semantic understanding helps.

When brainstorming or choosing an architecture, consider more than classification. The patterns below are starting points: combine primitives around the user's goal, including ideas that do not fit an established recipe.

  • Route and fill known arguments. A request can select a handler and its typed parameters. Ask useful branch-specific questions up front and consume only the relevant answers. See function calling and speculative fan-out.
  • Select instead of generate. Find candidate values or source spans in code, use a judgment to select the intended one, then copy or normalize it. Code can also assemble source text into a formatted document or reading guide. See value extraction and structure recovery.
  • Find and judge evidence. Retrieve candidates, compare their relevance to a query, and select useful context. See reranking and hierarchical classification.
  • Turn judgments into reusable data. Score dimensions once, then let code or user controls change weights, thresholds, rankings, and views. With labeled outcomes, those signals can become classical ML features. See composite scoring and feature discovery.
  • Verify and escalate. Check specific claims or fields against their evidence; send uncertain or failing cases to a person or reasoning model. See citation checks and extraction cascades.
  • Respond to changing state. Code can retain goals and observations while fresh judgments guide the next bounded step. Keep inferred state distinct from observed facts, and check freshness before applying a result to a changed situation.

For open-ended requests, offer the few directions that best serve the user's goal and recommend a starting point. For a concrete request, choose the relevant pattern and build; a brainstorm is not a mandatory detour.

Design the judgments

System One models use three primitive types. Choose by what the answer means:

NeedPrimitiveImportant distinction
One of a defined setChoicePicks one option; its probability distribution compares competing options
Whether a condition holdsNoulProbability of yes; no separate confidence; use one per label when several may apply
Degree along a described dimensionScoreProbability-weighted position on ordered levels; use comparable per-item Scores for graded ranking

Give each question enough relevant state to answer: source text, identities, relationships, policies, and current facts. Prefer named JSON fields when context has several parts. Put the judgment in instructions and define its possible answers in criteria. Question IDs are for code and are not sent to the model; include complete meaning in the question. Reference nested state with backticked paths such as ticket.messages[0].text.

Ask one narrow, coherent judgment per question. Split independently useful dimensions, without destroying the relationship being judged. A bounded action selection or contextual interpretation is valid; atomic does not mean literal fact extraction or a one-sentence limit. Strings work for simple questions. Use structured objects or arrays when definitions, contrasts, exclusions, or examples clarify instructions or criteria. Score levels must describe concrete situations and stand on their own.

Keep the needed answers available. Include a no-match outcome when nothing may fit; use a separate presence judgment when it is independently useful. For source-value selection, check candidate coverage: the model cannot choose an omitted value.

Compose and verify

Ask independent questions over the same state together, including useful speculative questions. They run in parallel and cannot see one another's answers. State each speculative premise explicitly; code consumes the applicable answers. A second request is warranted when an earlier answer is needed to fetch evidence, construct new state, or determine the next options. Extra questions still use tokens; measure actual request budgets, cost, and end-to-end latency.

Use probabilities and confidence to guide behavior, with thresholds evaluated on the user's data and consequences. Choice/Score confidence summarizes distribution concentration, not overall workflow correctness or permission to act. A Noul near 0.5 means similar probability for yes and no, not medium intensity. Several acceptable alternatives can also spread probability; low confidence need not invalidate a harmless preference choice. Ignore uncertainty on unused branches.

Keep policy explicit and raw judgments reusable. Weighted scores suit compensating preferences; an "any serious violation" rule needs separate conditions. Changing a weight or display filter need not rerun inference when evidence and question meanings are unchanged. Typed output guarantees the interface, not truth. System One models are trained for calibrated decisions; validate their performance in the target domain.

For patterns combining System One models with LLMs, see references/llm-integration.md — the five cascade patterns (gating, candidate selection, output verification, routing, ensemble features) cover the most common architectures.

Test representative cases and the resulting application behavior. For failures, inspect the exact state, questions, candidates, answers, composition, and observed outcome. Separate missing evidence, model errors, code errors, and service failures. Treat cookbook thresholds and demo results as examples to evaluate, not universal rules or permanent model limitations. Keep API credentials server-side in web apps.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.