File size: 3,122 Bytes
2eea493 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 | # GDPR — Privacy & Data Processing
How mona.expert processes personal data, and how the open-source client
minimises it. This document supports both end-user transparency
(Art. 13/14) and customer due diligence (Art. 28 annex-ready).
## Roles
- **mona.expert** — data controller for account data; data processor
where customers operate the platform on their own behalf.
- **You (the user)** — controller of your own device and of the
commands you issue.
## Processing activities (Art. 30 summary)
| Purpose | Categories | Legal basis | Retention |
|---|---|---|---|
| Account & authentication | Email, username, API token, session | Contract (Art. 6(1)(b)) | Account lifetime + 30 days |
| Agent operation & audit | Task text, tool results, audit events, conversation | Contract / legitimate interest (security) | 180-day rolling audit; conversation until deleted |
| Device telemetry | Hostname, OS, CPU/memory/disk/load metrics, IP | Contract (service operation) | Latest snapshot + 180 samples (~30 min) |
| Security & abuse prevention | IP, rate-limit counters, timestamps | Legitimate interest (Art. 6(1)(f)) | 24 h (rate limits), 180 days (audit) |
## Data minimisation by design
- **The device stores no provider keys.** Only a mona.expert token
lives locally (`~/.mona-agent/credentials.json`, mode 0600).
- **Telemetry is system metrics only** — no keystrokes, no screen
content, no file contents. Command results are sent only because you
asked the agent to run them.
- **Egress-only networking** — the daemon opens no inbound ports and
sends nothing to third parties.
## Security of processing (Art. 32)
- AES-256-GCM encryption at rest for all key material
- HTTPS/TLS for all traffic
- Per-user bearer tokens + session auth + CSRF protection
- Least-privilege tool sandbox, allowlisted shell
- Full audit trail of every action
## Data subject rights
- **Access / rectification / erasure** — the dashboard shows your
agents, conversations and audit entries; delete them there or write to
`privacy@mona.expert`.
- **Portability** — conversations and settings export via API.
- **Objection / restriction** — stop the agent; processing stops
(device disconnects, no further collection).
## International transfers
- Processing and storage occur in the EU (Hetzner/Hostinger EU data
centres via the mona.expert infrastructure). No data is transferred to
third countries by mona.expert itself.
- AI provider calls (OpenAI, Anthropic, Google, …) are made **with your
own keys**, under the terms you hold with those providers.
## Sub-processors
| Sub-processor | Purpose |
|---|---|
| Hosting (EU) | Infrastructure for the mona.expert cloud |
| AI providers (your keys) | Model inference on your instruction |
## DPIA note
mona-agent processes telemetry and command output, not special-category
data. For typical use a DPIA is not required; a template is available on
request for enterprise deployments (`compliance@mona.expert`).
## DPA
A data-processing agreement annex (Art. 28) is available for business
customers on request: `compliance@mona.expert`.
|