Skip to main content

EchoHuman Operating Contract

Status: Active

Purpose​

This contract governs how humans and AI agents work across the EchoHuman ecosystem.

The objective is to reduce drift, avoid unnecessary complexity, preserve accumulated knowledge and ensure that implementation follows evidence rather than assumption.

Core Rule​

Discover → Report → Review → Design → Approve → Implement → Test → Prove → Next Stage

No stage is skipped.

Discovery First​

Before any implementation:

  • inspect existing systems
  • inspect existing documentation
  • inspect existing infrastructure
  • inspect existing data models
  • identify reuse opportunities

Output: Discovery Report

No changes permitted.

No Changes Yet​

During discovery:

  • no deployments
  • no commits
  • no pushes
  • no migrations
  • no schema changes
  • no service restarts
  • no infrastructure modifications

Design Phase​

Produce:

  • architecture proposal
  • schema proposal
  • risks
  • rollback plan
  • implementation plan

Wait for approval.

Implementation Phase​

Build only approved scope.

Avoid:

  • scope creep
  • speculative architecture
  • unnecessary infrastructure
  • duplicate systems

Test Phase​

Prove:

  • functionality
  • data integrity
  • provenance integrity
  • rollback capability

Documentation Rule​

No major system exists properly until documented.

Documentation is part of the deliverable.

AI Agent Rules​

Applies to:

  • ChatGPT
  • Devin
  • Windsurf
  • Claude
  • future agents

Agents must:

  • inspect before modifying
  • document before building
  • prove before claiming success
  • preserve provenance
  • reuse existing systems

Evidence takes precedence over confidence.