Data governance
The datagov agent assesses the tenant’s registered data sources. It is deterministic and calls no model: a model-authored governance score would be a fabrication wearing a number. It holds no credential, dials no source, and reads no row or cell value.
When it runs
- daily at 07:11 UTC;
- on an
@tide_datagovissue comment; - on a manual run of the agent’s workflow;
- when an Atlas playbook step names
datagov.
Arm or disarm it with the gate agentic_datagov from DataEx › Agents.
What it reads
- The connection snapshot: a copy of the connection table that the console publishes after every write. This covers the governance and risk-compliance axes.
- What Resolve’s Explore recorded, read from the console with the agent’s own app token. It holds structure and counts only: names, declared types, nullability and primary keys, row and null counts with their basis, the source’s PII flag and mask, and access counts. This covers the data-quality and PII-exposure axes.
The agent reads it from the console on purpose, rather than from a published copy: a description is the tenant’s schema, so it stays sealed at rest in the console.
Verdicts
pass, fail, warn and unable-to-assess. The last one is the point: a control the agent cannot evaluate is never shown as satisfied or omitted, and it always carries the reason.
Controls
The agent groups its controls into four axes. Each control gives one verdict per connection, with the reason:
| Axis | What it covers |
|---|---|
| governance | whether the connection says who owns it and which environment it belongs to, whether its type is one the connector catalog describes, and whether its credentials are its own and kept out of its ordinary settings |
| risk-compliance | whether the connection declares where it reaches, keeps credentials out of that address, and protects the transport |
| data-quality (from Resolve’s Explore) | whether what a source declares about its own structure agrees with what was measured, and whether its tables are keyed |
| pii-exposure (from Resolve’s Explore and access patterns) | whether columns that may hold personal data are masked, and how widely such data can be read |
A source nobody has explored gets one unable-to-assess per axis, saying “not explored yet — run Explore in Intelligence › Resolve” (the page is now named Estate). To fill the two data axes, Explore the source on Estate.
The personal-data findings are a starting point for review, not a guarantee that every sensitive column was found.
Where the results land
- The console findings store: the agent writes its findings to the console with its app token, and signed-in users read them there. Findings name column paths so an owner can act.
- The living GitHub issue, “🗃️ Data governance status”. It carries counts only, so the tenant’s schema stays on the platform.
- A successful store write raises the harness event
findings.reported(fail_count,warn_count), which can start a matching Atlas playbook. See Playbooks.
Tests hold the value boundary: no finding carries a setting value or a cell value, and what Explore hands the agent carries no values.
Not the same as
compliance: checks Runink TIDE’s own code against its control index.risk: govulncheck over dependencies.judge: judges findings submitted by other platforms. Its docket is on the judge’s page in DataEx › Agents.