Repository navigation
hitl_governance
iso_doc_type: "Policy" iso_viewpoint: "SecurityView" type: "security" title: "HITL Governance β Dark Gravity CA/CD Autonomous Factory" description: "ISO 42010 SecurityView / ISO 15289 Policy documentation for HITL Governance β Dark Gravity CA/CD Autonomous Factory." tags: ['iso42010', 'okf', 'security_view', 'zero_trust', 'governance'] timestamp: "2026-10-01T14:00:00Z" generated: "agent:okf-professional-documenter" verified: "true" last_verified_commit: "fbdc5848"
Purpose: Formal specification of the 4-Vertex Human-in-the-Loop governance mesh ensuring human oversight at critical decision points in the autonomous pipeline.
The Dark Gravity Factory implements a 4-Vertex HITL Mesh β four strategically placed human decision gates that prevent fully autonomous operation while maximizing agent acceleration.
Core Principle: Agents are empowered to create branches, commit code, push upstream, create PRs/MRs, and resolve CI/CD failures. Under no circumstances may an agent automatically merge a PR or MR into
mainor default branches. The merge process MUST be a manual human action.
graph LR
V1["Vertex 1:<br/>Strategic Injection<br/>(Product Owner)"]
V2["Vertex 2:<br/>Sprint Mobilization<br/>(Tech Lead)"]
V3["Vertex 3:<br/>Exception Override<br/>(Architect)"]
V4["Vertex 4:<br/>Categorical Scrutiny<br/>(Senior Reviewer)"]
V1 -->|"Epic approved"| AGENT_PLAN["Agent Planning"]
AGENT_PLAN -->|"Tasks decomposed"| V2
V2 -->|"Tasks approved"| AGENT_CODE["Agent Coding"]
AGENT_CODE -->|"Deadlock detected"| V3
V3 -->|"Override applied"| AGENT_CODE
AGENT_CODE -->|"PR/MR created"| V4
V4 -->|"Merge approved"| DEPLOY["Deployment"]
style V1 fill:#4CAF50,stroke:#2E7D32,color:#fff
style V2 fill:#2196F3,stroke:#1565C0,color:#fff
style V3 fill:#FF9800,stroke:#E65100,color:#fff
style V4 fill:#9C27B0,stroke:#6A1B9A,color:#fff
| Attribute | Value |
|---|---|
| Human Role | Product Owner |
| Trigger Condition | Creation of Epic tagged autonomous-plan / autonomous-mission / dark-gravity
|
| Enforcement Mechanism | GitHub/GitLab label requirements; Poller only ingests issues with required labels |
| Decision | Whether a mission should be autonomously executed |
| Output |
PolledIssueEvent dispatched to ingestion pipeline |
| Attribute | Value |
|---|---|
| Human Role | Tech Lead |
| Trigger Condition | RustantAgent generates SddMissionPlan with decomposed tasks |
| Enforcement Mechanism | Hatchet DAG Plan phase waits for approval signal; /speckit-tasks generates tasks requiring human review |
| Decision | Whether the task decomposition is correct and safe to execute |
| Output | Approved SddTaskItem list dispatched to Code phase |
| Attribute | Value |
|---|---|
| Human Role | Architect |
| Trigger Condition | Aethelgard circuit breaker trips: 3 consecutive failures with stagnant diff hash |
| Enforcement Mechanism |
Agent-Stuck state triggers Slack/Jira alert; DAG pauses until human override |
| Decision | Override parameters, abort mission, or reassign to different agent strategy |
| Output | Resumed execution with tuned parameters OR mission abortion |
| Attribute | Value |
|---|---|
| Human Role | Senior Reviewer |
| Trigger Condition | PR/MR created by agent after successful validation + review phases |
| Enforcement Mechanism | Branch protection rules require human approval; agents cannot merge |
| Decision | Final design review, security posture validation, merge approval |
| Output | Merged PR/MR β FluxCD deployment |
stateDiagram-v2
[*] --> EpicCreated: PO creates tagged Epic (Vertex 1)
EpicCreated --> Ingested: Poller detects required labels
Ingested --> Planning: RustantAgent plans mission
Planning --> TasksReady: SddMissionPlan generated
TasksReady --> HumanApproval_V2: Vertex 2 - Tech Lead Review
HumanApproval_V2 --> Coding: Tasks approved
HumanApproval_V2 --> Rejected: Tasks rejected
Coding --> Validation: ZeroClaw executes TDD
Validation --> Review: Tests pass + SAST pass
Validation --> AgentStuck: 3 retries exhausted
AgentStuck --> HumanOverride_V3: Vertex 3 - Architect
HumanOverride_V3 --> Coding: Override with new params
HumanOverride_V3 --> Aborted: Mission rejected
Review --> PRCreated: RustantAgent creates PR/MR
PRCreated --> HumanMerge_V4: Vertex 4 - Senior Reviewer
HumanMerge_V4 --> Delivered: PR merged (human action)
HumanMerge_V4 --> Coding: Changes requested
Delivered --> [*]
Rejected --> [*]
Aborted --> [*]
The wiki documentation is synchronized via the .github/workflows/docs-to-wiki.yml pipeline:
sequenceDiagram
participant Dev as Developer/Agent
participant Git as Git Repository
participant CI as GitHub Actions
participant Wiki as GitHub Wiki
Dev->>Git: Push to wiki/ directory
Git->>CI: Trigger docs-to-wiki.yml
CI->>CI: Validate markdown links
CI->>CI: Check Mermaid syntax
CI->>CI: Calculate OSR (< 5% gate)
alt OSR < 5%
CI->>Wiki: Sync wiki/ β GitHub Wiki
CI-->>Dev: β
Wiki updated
else OSR >= 5%
CI-->>Dev: β OSR gate failed
end
Related: Security Architecture Β· Business Context Β· Compliance & Audit
- Business Context
- Strategic Design
- Tactical Design
- Agent Specifications
- Runtime Sequences
- Infrastructure Adapters
- Mission Data Model
- lib.md
- gitlab_verifier.md
- poller_service.md
- telemetry_export.md
- Agents
- Workflows
- Bridge
- Utils
- lib.md
- github.md
- gitlab.md
- jira.md
- kafka.md
- ziti.md
- vault.md
- r2r.md
- semantica.md
- aethalgard.md
- mcp_client.md
- pipeline_classifier.md
- s3.md
- sentry.md
- security_validator.md
- cursor_store.md
- git_poller.md
- lib.md
- main.md
- protocol.md
- sandbox.md
- scratch.md
- feedback_route.md
- github_webhook.md
- Skills
-
Tools
- mod.md
- bridge.md
- deep_research_tool.md
- execute_code.md
- get_factory_status.md
- index_code.md
- inspect_kafka_topic.md
- launch_sandbox_pod.md
- list_minio_buckets.md
- list_minio_objects.md
- plan_mission.md
- retrieve_context.md
- run_tests.md
- search_jira.md
- security_review.md
- spec_kit_tasks_to_issues.md
- spec_kit_tool.md
- update_mission_status.md