Framework roles
Founding steward, maintainers, contributors, independent reviewers, and affected-person reviewers govern public framework changes.
Who stewards it
Governance & About.
Life First separates governance of the public framework from accountability for any organization or system that uses it.
Framework governance
Independent stewardship stays explicit.
Life First began with Roshi's proposal for AI practice that actively protects life while preserving dignity, agency, lawful process, and accountability. It remains an independent project stewarded by Roshi. Sentinel AI assists drafting, implementation, verification, and repository operations.
Traceable changes
Substantive changes retain rationale, safeguard impact, provenance, reviewers, validation, compatibility impact, and review dates.
Deployment accountability
Using the framework does not transfer responsibility.
Every Tier 1–3 workflow names the people with real authority, competence, time, information, and obligation to own, challenge, stop, correct, and answer for the decision.
Accountable owner
Owns purpose, resources, residual risk, and the decision to operate or stop.
Safety and domain leads
Own hazards, controls, thresholds, validity, professional boundaries, and external obligations.
Privacy and operations
Own data necessity, access, retention, deletion, least privilege, continuity, rollback, and emergency stop.
Review and remedy
Own reviewer conditions, overrides, explanations, appeals, correction, repair, and affected-person challenge.
Connection to The Wizard Nexus
Connected does not mean owned.
Life First is independently stewarded by Roshi. The Wizard Nexus describes it as a human-first governance framework that informs consequential work across its ecosystem. TWiN does not own, certify, or govern Life First.
Life First
Authority, proportionality, agency, safeguards, accountable action, and repair.
Open Life First →TWiN Compass
Ethical and behavioral-health-informed orientation for purpose-built AI.
Visit TWiN Compass ↗KEMPO
Evaluation, testing, human confirmation, monitoring, and learning.
Visit KEMPO ↗Astrolabe
A public map of where relationships and influence exist across the TWiN portfolio.
Visit Astrolabe ↗Provenance and acknowledgment
Influence is visible. Boundaries stay explicit.
Dialogue with Chris Blask — QuietWire and Lumina — Civic AI, QuietWire, together with public Civic AI Canon materials, informed attention to semantic attestation, auditability, narrative sovereignty, operational discipline, and constructive human-machine collaboration.
Life First retains its own repository, stewardship, development rhythm, and site. No contribution, acceptance, reciprocal link, endorsement, sponsorship, or affiliation is claimed without a specific public record.
Secondary resources
These already exist in the repository.
They were not all visible from the previous public page. They remain source documents for now while the basic page structure is established.
FAQ
Certification, urgency, human review, public data, scoring, non-AI alternatives, retirement, and project relationships.
Read FAQ ↗Glossary
Normative language, operational terms, accountability, agency, appeals, controls, repair, shadow mode, and stop conditions.
Read glossary ↗Changelog
Versioned public changes, additions, clarifications, and safeguard impact.
Read changelog ↗Security
Public and sensitive reporting routes, publication boundaries, and response priorities.
Read security guidance ↗GitHub source
The normative text, implementation documents, templates, schema, cases, governance, and version history.
Inspect source ↗Contribute
Bring a real use case and a willingness to learn.
Proposals should identify affected people, legitimate purpose, foreseeable misuse, safeguard impact, measurable benefit, counter-metrics, implementation impact, provenance, and reuse terms.