AGENTS.md — NeuroLift Technologies haief Repository Agent Gateway

Internal use only. This is the repository-specific governance gateway for coding agents operating in this repository. This copy is scoped to NeuroLift-Technologies/haief, with project-specific operational context kept in this repository.

You Are Here

You are a coding agent operating within the NeuroLift Technologies organization. This document is your repo-local coordination gateway.

Mandatory reading order:

  1. NLT-DEV-OTOI.md — Repo-local coding agent contract (this repo, root level)
  2. Repo-level CLAUDE.md — Project-specific context (in the repo you are working in)
  3. docs/active-threads.md — Current work state (in the repo you are working in)

Repo-local note: Governance docs in this repository are the working copy for NeuroLift-Technologies/haief. Keep references local unless Joshua explicitly asks for an upstream sync.

Final authority: Joshua W. Dorsey, Sr. Escalate. Do not guess.

Solidarity Framework Principles (Public)

The ethical foundation of all NLT work is publicly documented in the Solidarity Framework and HAIEF (Human-AI Ethical Integration Framework):

The principles are public. The operational machinery for this work lives in this repository.

Coordination Protocol

Session Start (Every Session)

1. Read NLT-DEV-OTOI.md (this repo)
2. Read repo-level CLAUDE.md (working repo)
3. Read docs/active-threads.md (working repo)
4. Self-register per OTOI Section 3
5. Confirm task scope before beginning

Commit Format

All agent commits must follow:

[AGENT_NAME] type(scope): description

Types: feat, fix, docs, refactor, chore, test, ci

Escalation Triggers

Escalate to Joshua immediately when:

Use the escalation template: templates/escalation.md

Guardrails

These are non-negotiable. No exceptions without explicit Joshua approval:

Guardrail Details
No LLM provider lock-in Do not hardcode or commit to a specific LLM provider
No architecture decisions Database, deployment, framework choices require human sign-off
No production deployments Human must explicitly approve all production actions
No credential storage Never store secrets, tokens, or credentials in code or VCS
No external integrations Third-party service connections require Joshua’s approval
No OTOI self-amendment This governance doc cannot be changed by agents

Internal File Map

All files below live in this repository (NeuroLift-Technologies/haief):

NLT-DEV-OTOI.md                        ← Repo-local agent contract
AGENTS.md                               ← This file
nltotoi.json                            ← Discovery manifest

.nltotoi/
├── README.md                           ← Namespace overview
├── index/governance-files.md          ← File registry
├── contracts/README.md                ← Contract namespace
├── proposals/validation-roadmap.md    ← Validation roadmap
└── scripts/validate-governance.sh     ← Governance validation

templates/
├── agent-registration.json            ← OTOI Section 3 registration format
├── handoff-record.json                ← OTOI Section 5 handoff format
├── escalation.md                      ← OTOI Section 4.3 escalation format
└── intent-log.md                      ← Intent logging template

ISSUE_TEMPLATE/
├── agent-escalation.md                ← GitHub escalation issue form
└── governance-proposal.md             ← OTOI amendment proposal form

PULL_REQUEST_TEMPLATE/
└── agent-contribution.md              ← Agent PR checklist

workflows/
└── validate-governance.yml            ← CI: runs validate-governance.sh

SOPs/
├── new-agent-onboarding.md            ← How to onboard a new coding agent
├── repo-governance-setup.md           ← How to add governance to a new NLT repo
└── incident-response.md               ← What to do when an agent goes off-rails

Multi-Agent Coordination

When multiple agents may be active:

  1. Check active-threads.md first — do not begin work already in progress
  2. Claim your thread — update active-threads.md when starting a task
  3. Write handoff records — never leave a session without a handoff document
  4. Do not overwrite peer work — if conflict is detected, escalate

Handoff Protocol

Before ending any significant session:

  1. Update docs/active-threads.md in the working repo
  2. Write a handoff record to docs/agent-log/handoffs/ using templates/handoff-record.json
  3. Document any open escalations in docs/escalations/
  4. Summarize decisions made and decisions pending
*Internal governance document — NeuroLift Technologies ORG-DEV-OTOI-1.0.0*