Jester v0.2 — Causal Market Imagination

Jester v0.2 — Causal Market Imagination

Jester v0.2 is a standalone, deterministic market-world research platform that connects point-in-time evidence, narrative diffusion, competing causal explanations, strategic participant behavior, price and volatility-surface paths, liquidity, and executable long-option economics.

Headline systems

• Jester Observatory: a responsive read-only 3D cockpit with World, Options, Meta, Dream, and Sources modes.
• Dream Foundry: 24 typed cognition operators, four oneiric phases, recursive lineage, cognitive immunity, quality-diversity archives, and preserved failure memory.
• Scientific Literature Lab: autonomous versioned arXiv metadata discovery converted into falsifiable research capsules with zero automatic empirical credit.
• Attention Observatory: policy-admitted, robots-aware search-prominence metadata snapshots and chronological attention deltas.
• Market-world simulation: 10 simulator-family contracts, 8 market packs, 13 participant classes, and a Level 0–7 proof ladder.
• Evidence and intelligence mesh: 15 permissioned pipelines and 13 composable intelligence services.

Proof boundary

The included public fixture certifies deterministic mechanics through synthetic Levels 0–5. Forward-shadow and bounded-paper levels remain blocked without real forward evidence. Jester has no broker credentials, account surface, order construction, or execution authority.

Release verification

• 37 tests pass on Python 3.12 and 3.13
• Ruff format and lint pass
• strict mypy passes across 59 source files
• Python wheel and source distribution build
• frontend TypeScript and production build pass
• public safety and credential-material scans pass
• desktop and mobile browser QA pass with no console errors or horizontal overflow

Explore the public showcase: https://obtuseai.github.io/jester/

Jester v0.1.0 - Earnings Pack vertical slice

Jester v0.1.0 is the first deterministic, paper-only Earnings Pack vertical slice.

Highlights:

• immutable point-in-time evidence and governed mixed-source intake
• causal scenario lattices, numerical market society, and strategic game stresses
• executable long-option outcome fields with holdability, fill, IV, and surface checks
• correlation-aware belief exchange and bounded recursive hypothesis evolution
• independent proof manifests and Ed25519 governor candidate bundles
• bounded supervision and a local read-only observatory

The included public fixture is synthetic and resolves only as mechanics_certified / certified_mechanics_only. This release makes no predictive-edge, market-fidelity, forward-performance, or state-of-the-art claim.

Validation: Python 3.12 and 3.13 CI, strict mypy, Ruff, 17 tests, package build, and public-safety scan.