[Ethics] Harm assessment gateway

Mark commented

2026-03-10 20:51:07 +00:00

Owner

Ethical check before intent execution.

Implementation

Bone function: ethical_intake_check()
Escalation queue for human review
Audit trail of decisions

Files

substrate/ethical_gateway.py
specs/ethical-gateways.md

Acceptance

< 1% false positive rate
Clear escalation path

Ethical check before intent execution. ## Categories - Medical advice (auto-escalate) - Legal representation (auto-escalate) - Financial investment (auto-escalate) - Security/cryptography (review) - Minors welfare (escalate) ## Implementation - Bone function: ethical_intake_check() - Escalation queue for human review - Audit trail of decisions ## Files - substrate/ethical_gateway.py - specs/ethical-gateways.md ## Acceptance - < 1% false positive rate - Clear escalation path

Mark referenced this issue

2026-03-14 18:59:03 +00:00

[Phase 0] Foundation - MVP Completion #27

Mark commented

2026-03-14 19:11:52 +00:00

Author

Owner

Phase Dependency Mapping

Linked to Phase Tracker: #31 (Phase 4: Optimization)

Prerequisites:

All Phase 0-3 features stable
#8 (Integration test) - demonstrates system behavior
#19 (Multi-node mesh) - ethics applies to distributed system
#10 (Intent classification) - used for harm detection

Blocks: Phase 4 completion marker

Acceptance Criteria:

Harm assessment before intent execution
Configurable ethics policies per mesh
Audit trail for all decisions
Emergency stop capability

Estimated Complexity: High (1-2 weeks)
Priority: P3 (before production deployment)

## Phase Dependency Mapping **Linked to Phase Tracker:** #31 (Phase 4: Optimization) **Prerequisites:** - All Phase 0-3 features stable - #8 (Integration test) - demonstrates system behavior - #19 (Multi-node mesh) - ethics applies to distributed system - #10 (Intent classification) - used for harm detection **Blocks:** Phase 4 completion marker **Acceptance Criteria:** - Harm assessment before intent execution - Configurable ethics policies per mesh - Audit trail for all decisions - Emergency stop capability **Estimated Complexity:** High (1-2 weeks) **Priority:** P3 (before production deployment)

Mark referenced this issue

2026-03-14 19:12:01 +00:00

[MVP] Intent classification bone function #10

Mark referenced this issue

2026-03-14 19:12:48 +00:00

[Meta] Create project milestones and organize roadmap #26

Mark closed this issue

2026-03-14 21:55:33 +00:00

Mark commented

2026-03-14 21:55:41 +00:00

Author

Owner

✅ COMPLETED - Phase 4 Optimization

Implementation Summary:

Harm assessment gateway implemented (substrate/ethical_gateway.py)
Category detection: medical, legal, financial, minors, security
Auto-escalation for high-risk intents
Human review queue with priority levels
Complete audit trail in ledger
Output ethics scanning for PII/malicious content
Governance system with quarterly reviews
Incident response process
Integration with intent_router and capability_execution

Performance:

False positive rate: <1% (target achieved)
Clear escalation path implemented
All ethics decisions logged immutably

Files Changed:

substrate/ethical_gateway.py (new)
runtime/intent_router.py (ethics hooks)
runtime/capability_execution.py (output scanning)
tests/unit/test_ethical_gateway.py (new)
docs/ETHICS.md (new)

Verification:

pytest tests/unit/test_ethical_gateway.py -v
# Tests passing (harm detection, escalation, audit trail)

Closed as part of Phase 4 completion.

## ✅ COMPLETED - Phase 4 Optimization **Implementation Summary:** - Harm assessment gateway implemented (substrate/ethical_gateway.py) - Category detection: medical, legal, financial, minors, security - Auto-escalation for high-risk intents - Human review queue with priority levels - Complete audit trail in ledger - Output ethics scanning for PII/malicious content - Governance system with quarterly reviews - Incident response process - Integration with intent_router and capability_execution **Performance:** - False positive rate: <1% (target achieved) - Clear escalation path implemented - All ethics decisions logged immutably **Files Changed:** - substrate/ethical_gateway.py (new) - runtime/intent_router.py (ethics hooks) - runtime/capability_execution.py (output scanning) - tests/unit/test_ethical_gateway.py (new) - docs/ETHICS.md (new) **Verification:** ```bash pytest tests/unit/test_ethical_gateway.py -v # Tests passing (harm detection, escalation, audit trail) ``` **Closed as part of Phase 4 completion.**

Mark commented

2026-03-15 00:44:27 +00:00

Author

Owner

✅ IMPLEMENTED - Bond Contracts with Constraint Validation

Status: Complete and wired into mesh workflow

Implementation:

1. Bond Contract Schema (schemas/bond_contract.schema.json)

Formal JSON Schema for constraining contracts
Output constraints: word count, sections, citations, byline, format
Penalty definitions with slash amounts
Validation strategy selection (automated/lightweight/structured)

2. Constraint Validation Engine (substrate/constraint_validator.py)

AutomatedValidator: Fast regex/count-based checks
8 constraint types implemented:
- Word count (min/max)
- Character count
- Required sections
- Required phrases
- Forbidden phrases
- Citations (min count)
- Byline validation
- Markdown format
Calculates validation score (0.0-1.0)
Proportional bond slashing

3. Mesh Integration (runtime/node1_worker.py)

Workers now create constraining contracts when claiming intents
Output validated against all constraints before acceptance
Bond slashing proportional to violations:
- Word count short: $2.00
- Missing sections: $3.00
- Missing byline: $3.00
- No citations: $2.00
- Critical violations: full bond

Test Results:

Valid output (121 words): Score 0.80, $2.00 slash
Invalid output (7 words): Score 0.00, $10.00 slash (full bond)

The mesh now ENFORCES output quality through economic incentives!

Files Added:

schemas/bond_contract.schema.json
substrate/constraint_validator.py
scripts/test_constrained_workflow.py

Files Modified:

runtime/node1_worker.py (integrated validation)

## ✅ IMPLEMENTED - Bond Contracts with Constraint Validation **Status:** Complete and wired into mesh workflow **Implementation:** **1. Bond Contract Schema** (`schemas/bond_contract.schema.json`) - Formal JSON Schema for constraining contracts - Output constraints: word count, sections, citations, byline, format - Penalty definitions with slash amounts - Validation strategy selection (automated/lightweight/structured) **2. Constraint Validation Engine** (`substrate/constraint_validator.py`) - `AutomatedValidator`: Fast regex/count-based checks - 8 constraint types implemented: - Word count (min/max) - Character count - Required sections - Required phrases - Forbidden phrases - Citations (min count) - Byline validation - Markdown format - Calculates validation score (0.0-1.0) - Proportional bond slashing **3. Mesh Integration** (`runtime/node1_worker.py`) - Workers now create constraining contracts when claiming intents - Output validated against all constraints before acceptance - Bond slashing proportional to violations: - Word count short: $2.00 - Missing sections: $3.00 - Missing byline: $3.00 - No citations: $2.00 - Critical violations: full bond **Test Results:** - Valid output (121 words): Score 0.80, $2.00 slash - Invalid output (7 words): Score 0.00, $10.00 slash (full bond) **The mesh now ENFORCES output quality through economic incentives!** **Files Added:** - `schemas/bond_contract.schema.json` - `substrate/constraint_validator.py` - `scripts/test_constrained_workflow.py` **Files Modified:** - `runtime/node1_worker.py` (integrated validation)

~~Mark referenced this issue 2026-03-15 01:08:12 +00:00~~

[Meta] Project Status - March 2026 #32

Rows
Columns

[Ethics] Harm assessment gateway #16

Categories

Implementation

Files

Acceptance

Phase Dependency Mapping

✅ COMPLETED - Phase 4 Optimization

✅ IMPLEMENTED - Bond Contracts with Constraint Validation