Codex Science v0.6.0 — Visual Research Workspace

Codex Science v0.6.0 adds a self-contained visual research workspace and durable cross-session handoff.

Highlights:

• .science/PORTAL.html unifies workflow stages, next actions, evidence, datasets, experiments, artifacts, providers, services, and capability gaps.
• .science/RESUME.md carries the study into a new Codex thread or another computer.
• .science/PARITY.json audits public Claude Science feature coverage as ready, degraded, or unavailable.
• science.py portal, resume, and parity expose the new experience through the unified CLI.
• Handoff packets now include continuity and public-capability context.
• Portal output is script-free, self-contained, CSP-protected, HTML-escaped, and secret-tested.

Validation:

• 26 offline tests passed.
• Plugin, skill, release, and installed-cache checks passed.
• GitHub CI passed on Python 3.10 and 3.12.

Boundary: this is public-feature conformance, not a claim of model-quality parity, access to proprietary Claude Science services, or complete product equivalence.

Codex Science v0.5.0

One-stop research concierge

• Adds a single science.py entry point for service discovery, guided initialization, secret-free environment checks, workflow configuration, status dashboards, next actions, connectors, evals, loops, validation, audits, forks, and handoff.
• Changes the coordinating skill from a script menu into a research concierge that inspects the workspace, asks only material questions, executes available work, and reports outcome, evidence, uncertainty, and the next falsifiable action.
• Adds Quick, Standard, and Deep profiles with domain metadata and explicit required, adaptive, or not-requested stages.
• Integrates .science/STATUS.json and .science/STATUS.md into project validation and research packets.
• Allows draft packets with known gaps but refuses to claim milestone completion while required stages or audit findings remain open.
• Adds a machine-readable 11-service catalog and safe credential-presence checks without persisting values.

Verification

• 24 offline unit and integration tests.
• CI passes on Python 3.10 and 3.12.
• All 8 skills, 22 plugin scripts, manifests, JSON files, and secret scans pass release validation.
• Installed-cache smoke passed, including a live bounded Crossref query and an intentionally blocked incomplete handoff.

Workflow coverage is navigation metadata, not a scientific-quality score, ethics approval, safety determination, expert review, or publication decision.

Codex Science v0.4.0

Highlights

• Adds read-only Crossref and PubMed metadata connectors plus keyed OpenAlex support.
• Preserves sanitized provider snapshots and imports only explicitly selected metadata records.
• Adds an eight-task transparent scientific-agent benchmark with immutable outputs, deterministic grading, optional blinded human scoring, and same-suite comparison.
• Integrates connector snapshots and evaluation runs into project validation and research packets.
• Hardens credential routing, redirect handling, release validation, tests, and Node 24 GitHub Actions.

Verification

• 19 offline unit and integration tests.
• CI passes on Python 3.10 and 3.12.
• Live read-only Crossref and PubMed smoke tests passed.
• OpenAlex support is implemented but requires a user-owned OPENALEX_API_KEY; no live OpenAlex run is bundled.

The bundled benchmark supports regression and exploratory same-task comparison. It is not proof of overall scientific intelligence or complete Claude Science parity.

Codex Science v0.3.0 — Loop Engine

Adds a bounded plan → trace → eval → decision research improvement loop; explicit success, iteration, budget and stall gates; pinned third-party capability registry; static skill/plugin risk scanning with immutable checkout verification; automatic loop validation from the main project validator; loop evidence in research packets; and 12 lifecycle, negative, integrity and migration tests. External capabilities remain opt-in, revision-pinned and subject to separate runtime authorization.

Codex Science v0.2.0

First public release of the auditable Codex Science workbench. Includes coordinator and specialist skills, append-oriented research ledgers, dataset and artifact hashing, approval-aware compute planning, study forks, independent review protocols, deterministic audits, research packets, migration support, CI, and remote-install validation.