/home/techb158/cosmic-risk.abdallabala.com/docs
Edit: /home/techb158/cosmic-risk.abdallabala.com/docs/dashboard-spec.md (3691B)
# COSMIC AI-Risk Dashboard Specification
## Purpose
The dashboard operationalizes COSMIC AI-Risk Management as a software artifact. It helps a project team quantify, review, mitigate, and govern AI project risks.
## Source-derived foundation
The implementation uses these source-derived concepts:
1. AI projects require a new risk approach because they depend on data quality, model stability, experimentation, opacity, effort uncertainty, and deployment challenges.
2. The research question focuses on quantifying AI project risks with a standardized metrological framework that integrates organizational, technical, and human dimensions.
3. The conceptual foundation includes COSMIC, ISO 19761, ISO 15939, SWEBOK, VIM, and Zaouache's governance triangle.
4. COSMIC-Risk aims to define indicators, build a software prototype, create an API, and validate the framework.
5. The governance triangle structures risk into organizational, technical, and human dimensions.
## Implementation extensions
The source deck does not define a concrete dashboard scoring engine. This package proposes the following implementation logic:
1. Risk score = probability x impact x detectability, normalized to 0 to 100.
2. Residual score = normalized score reduced by mitigation progress and expected mitigation effectiveness.
3. Overall score = average residual active risk plus approval penalty for unresolved high and critical risks.
4. Gate status = Ready, Warning, or Blocked based on risk score, mitigation completeness, data readiness, model metrics, legal review, ethical review, and approval state.
## Dashboard views
| View | Purpose |
|---|---|
| Overview | Show risk score, gate status, governance triangle, lifecycle readiness, and top risks. |
| Risk register | Track all risks, owners, scores, status, approvals, and evidence. |
| Deployment gate | Apply measurable criteria to determine release readiness. |
| Indicators | Show measurands, units, and interpretation rules. |
| Experiments | Connect model performance evidence to risk governance. |
| Project-management mapping | Map the dashboard model to Trello, Jira, Asana, and Microsoft Planner work-management objects. |
| Source alignment | Separate source-derived concepts from implementation extensions. |
## Gate criteria
| Criterion | Rule |
|---|---|
| Overall risk score | Must be less than or equal to 50. |
| Open critical risks | Must be 0 unless approved or accepted. |
| Mitigation completeness | Must be at least 60 percent. |
| Data readiness | Must be at least 70. |
| Selected model F1 | Must be at least 0.80. |
| Selected model stability | Warning below 75. |
| Ethical review | Must be approved or accepted. |
| Legal review | Must be approved or accepted. |
| High-risk approval state | No unapproved high risks. |
## Project-management integrations
The prototype supports a shared integration model for Trello, Jira, Asana, and Microsoft Planner. Trello remains useful for the Power-Up prototype, but the dashboard should not be designed as Trello-only.
| COSMIC object | Trello | Jira | Asana | Microsoft Planner |
|---|---|---|---|---|
| Project | Board | Project | Project |
| Risk | Card | Issue | Task Task |
| Lifecycle phase | List | Workflow status or field | Section or custom field |
| Dimension | Label | Label or component | Tag or custom field |
| Risk score | Custom field or badge | Custom field | Custom field |
| Mitigation | Checklist item | Sub-task or linked issue | Subtask Checklist item |
| Evidence | Attachment or comment | Attachment or comment | Attachment or comment |
| Owner | Member | Assignee | Assignee |
| Due date | Card due date | Issue due date | Task due date |
