HUB_Optimus

Reality → Evidence → Inference → Narrative → Operational Signal

Integrity-first diplomatic simulation workflow for evaluation, preventive mediation, and systemic learning.

Languages / Idiomas:

Primary onboarding: EN · ES · DE

Translation status: CA / FR / RU = progressive translation (not language-faithful onboarding yet) · HE = stub / full translation pending · ZH = governance stub only

Source of truth: docs/context/STATUS.md

Quick paths:

• 90 seconds: EN · ES · DE
• 20 minutes (canonical v1 ES): v1_core/languages/es
• Hands-on: Scenario Template

---

What it is

HUB_Optimus is developed in a publicly visible repository with restricted rights; contribution and use are governed by IP_NOTICE.md. It is designed to improve diplomatic outcomes through:

• structured evaluation (incentives, verification, sequencing),
• preventive, non-coercive mediation options,
• iterative meta-learning from both failures and successes.

It helps humans and institutions:

• avoid repeating known historical failure patterns,
• detect “false successes” before they escalate,
• align incentives with medium/long-term stability,
• operate rationally under time pressure and political noise.

What it is not

• Not an authority.
• Not a prediction engine.
• Not a replacement for diplomacy.

It is a tool for better judgment.

---

Start in 60 seconds

New here?

• Start (EN): docs/00_start_here.md
• Empezar (ES): docs/es/00_start_here.md
• Start (DE): docs/de/00_start_here.md

See it in practice (guided walkthrough):

• 🇬🇧 Try a scenario: docs/03_try_a_scenario.md
• 🇪🇸 Probar un escenario: docs/es/03_try_a_scenario.md

Go deeper (workflow / simulator):

• 🇬🇧 Workflow: v1_core/workflow/README.md
• 🇪🇸 Workflow (ES): v1_core/workflow/es/README.md

---

Why this project exists

Modern diplomatic and institutional systems often fail not because of lack of intelligence, but because of:

• misaligned incentives,
• short-term optics overriding long-term stability,
• absence of early correction mechanisms,
• ignored historical recurrence.

HUB_Optimus exists to break that cycle by evaluating scenarios before decisions become irreversible.

---

Core principles (high-level)

• Stability over optics
  Medium/long-term systemic stability is the supreme criterion.

• Integrity first
  Influence over the core is earned through ethical coherence, not position or credentials.

• Evaluation over narrative
  Outcomes are assessed structurally (incentives, verification, sequencing), not rhetorically.

• Prevention over reaction
  Early, discreet mediation is preferred to public escalation.

• No scapegoating
  Errors are treated as systemic, not personal.

---

Repository map

• docs/ → onboarding and reading paths (recommended entry point)
• v1_core/ → active kernel: architecture, operational flow, workflow, templates, scenarios, meta-learning
• docs/architecture/runtime_contract.md → full technical contract (schema, runtime, CI, encoding)
• legacy/ → historical/exploratory materials (v0), preserved for transparency

Source-of-truth policy is defined in docs/context/STATUS.md: v1_core/languages/es/ is canonical and v1_core/languages/en/ is parity reference.

---

The simulator

Current prototype (hub_optimus_simulator.py + run_scenario.py):

• loads and validates scenario files against a JSON schema,
• runs round-based negotiation with configurable actors and policies,
• checks simple offer-based success criteria,
• produces a deterministic JSON report (status, rounds, history, detail).

Framework design objectives (not yet implemented in the runtime):

• evaluate incentive alignment and sequencing,
• classify long-term stability risk,
• detect false successes (agreements that appear to succeed but carry systemic instability),
• propose non-coercive corrective options,
• support iterative meta-learning from historical scenario outcomes.

It does not predict outcomes. The current runtime evaluates whether a simple configurable condition is met. The broader structural evaluation capabilities are part of the project vision described in the governance and methodology documents.

Start here:

• v1_core/workflow/README.md

---

Contributing

See: CONTRIBUTING.md
Link-checking is enforced via GitHub Actions (Lychee).

---

Contact / Collaboration

If you want to collaborate (scenarios, methodology, review), open an issue or pull request including:

• intended use-case (training / policy review / research),
• target domain,
• constraints (time, verification, actors),
• desired outcome type (stabilizing / risk-reducing / de-escalatory).