Governance export
CFO-ready compliance data from the autonomous governance system. Every metric links to its source for independent verification.
Evidence-backed claims
This export follows the principle that every claim is backed by evidence a prospect can verify. Numbers shown here are derived from observable system state. Where data is not yet available, this is explicitly stated rather than estimated or fabricated.
Audit report counts
The auditor agent evaluates the live site against the published constraint set multiple times daily. Each audit produces a timestamped report stored in the repository.
| Metric | Value | How to verify |
|---|---|---|
| Total audit reports | 208 | Count files in /audit/ directory (excluding .gitkeep) |
| First audit report | 2026-05-02 | Oldest dated file in /audit/ |
| Most recent audit | 2026-05-18 | Most recent dated file in /audit/ |
| Average audits per day | 12.2 | 208 reports over 17 days |
Constraint pass/fail rates
The constraint set defines 44 machine-readable rules the site must satisfy. Each audit evaluates all constraints and records pass, fail, or skip status.
| Category | Constraints | Description |
|---|---|---|
| preservation | 4 | Content freshness and legal pages |
| visibility | 4 | Audit, tasks, agents, constraints published |
| honesty | 2 | Claims backed by evidence, commits attributed |
| reflexivity | 5 | Self-documenting system behavior |
| liveness | 3 | System actively operating |
| containment | 9 | Consistency and design standards |
| discoverability | 8 | SEO, structured data, feeds |
| progress | 9 | Roadmap delivery and milestones |
Current violations (as of latest audit):
- CST-049 (testimonial_widget_published): Awaiting verifiable client testimonials
- CST-017 (belgium_ai_news_preserved): Content freshness threshold exceeded
Full details: Latest audit report
Task completion metrics
Tasks are created by the auditor when it detects constraint violations, or by mue-builder-daily when roadmap items are ready. The dev-runner claims and completes tasks autonomously.
| Status | Count | Description |
|---|---|---|
| Done | 140 | Tasks completed by dev-runner-mue-site |
| In progress | 2 | Tasks currently claimed and being worked |
| Blocked | 9 | Tasks awaiting human input or clarification |
| Cancelled | 13 | Dev tasks cancelled (superseded or invalid) |
Each completed task represents autonomous work: code written, files committed, and deployed without a human developer in the loop. The decision log shows the full chain from violation detection to resolution.
Agent activity summary
Eight autonomous agents operate the governance system, each with a defined scope and schedule.
| Agent | Trigger | Purpose |
|---|---|---|
| mue-inbox | Hourly (09-21 UTC) | Classify and respond to incoming mail |
| mue-audit-hourly | Hourly (09-21 UTC) | Detect health issues with inbox agent |
| mue-audit-daily | Daily (18:00 UTC) | Send daily summary to Florian |
| mue-strategist-monthly | Monthly (1st, 06:00 UTC) | Keep roadmap aligned with mission |
| mue-translator-weekly | Weekly (Mon, 05:00 UTC) | Sync constraints with roadmap progress |
| mue-builder-daily | Daily (07:00 UTC) | Convert roadmap items to dev tasks |
| dev-runner-mue-site | Hourly | Execute dev tasks on the site repo |
| mue-site-auditor | Weekly (Mon, 06:00 UTC) | Evaluate site against constraints |
Each agent operates under a published charter specifying what it can and cannot do. See /agents/ for full details.
Cost impact metrics: not yet available
Calculating cost impact requires operational baseline data from a client engagement. This includes:
- Hours spent on manual compliance checks before automation
- Cost per violation remediation under manual process
- Time-to-fix comparison (manual vs autonomous)
- FTE allocation before and after deployment
This data will be populated as the system accumulates operational history with client engagements. The governance system tracks its own operational output; cost translation requires client-specific context.
See the ROI framework for how these metrics will be calculated when baseline data becomes available.
Export data sources
All governance data is accessible through these pages:
For CFO review
This export demonstrates that:
- Governance is observable. 208 audit reports, 44 constraints, 140 completed tasks, all timestamped and attributed.
- Compliance is continuous. The auditor runs 12+ times daily, not annually or quarterly.
- Violations are tracked to resolution. From detection in an audit report to fix shipped by the dev-runner, the full chain is documented.
- Agent behavior is bounded. Each of the 8 agents operates under a published charter specifying what it can and cannot do.
For questions about governance data or to discuss how this applies to your organization, get in touch.