# Veto (vetosql.com)

Checks SQL an AI agent plans to run and returns an ok, warn or block verdict with stable finding ids.

- Atlas card: https://inite.ai/en/atlas/vetosql.com
- Site: https://vetosql.com
- For: Teams whose AI agents write Postgres SQL and need a safety check before it runs
- Categories: data
- Use 0/100 · alive 100/100

## Used for

- Block unscoped deletes before an agent runs them
- Gate CI builds on SQL safety findings
- Flag lock-heavy migrations for review before running

## Price

Not read yet.

## For developers

- MCP server: com.vetosql/veto (namespace; open on 2026-10-25, 3 tools)
- Read 2026-10-25

## MCP server, measured daily

Server: https://vetosql.com/mcp, checked every day since 2026-10-25 (2 checks).

- Answered: measuring, 2 daily checks so far
- Median first answer: 394 ms
- Sign-in: none
- Context its tool definitions take: about 697 tokens (3 tools)
- Tool definitions: B, 78/100
- Failing "params_described": 1 of 3 tools
- Failing "annotations": 3 of 3 tools
- Warning "open_writes": set_policies
- Method: https://inite.ai/en/atlas/mcp/report

## Your data

- Certifications stated: gdpr
- Read 2026-10-25

---

Source: INITE Atlas, https://inite.ai/en/atlas/vetosql.com. JSON: https://inite.ai/api/atlas/vetosql.com. MCP: https://inite.ai/api/mcp (atlas_lookup).