monte-carlo-prevent
monte-carlo-data/mc-agent-toolkit / skills/prevent
checks the impact of a dbt model change , which improves: avoids breaking production
Details
- judge’s score
- 9/10
- place
- outside the packs
- publisher
- monte-carlo-data
- status
- new
- pinned commit
- f1ea90f on GitHub (new tab)
- skill’s last commit
- Jul 3, 2026
- repo stars
- 93
- main type
- Data
Judgement
Very clear triggering and precise rules (W1 to W5, markers, parameter tables). Score capped by inconsistent tool names (getTable vs get_table), a truncated appendix and a very imperative tone.
reservations
- MCP tool names inconsistent between SKILL.md (getTable, getAlerts) and the appendices (get_alerts, get_table), which may break calls.
- references/workflows.md is only provided truncated (5000 characters out of 36574): steps W2 to W5 cannot be fully assessed.
- Vendor-specific skill (Monte Carlo): useless without an MC account, and it forces routing to its MCP server even when the user configures another one.
- Triggering
- 2/2
- Added value
- 2/2
- Precision
- 2/2
- Scope
- 2/2
- Quality
- 1/2
Alerts
3 alerts, by category
Network
- high risk
Sends table metadata and query history to the third-party server mcp.getmontecarlo.com (OAuth), not disclosed in SKILL.md.
Execution
- high risk
Runs dbt build and SQL queries on the data warehouse: protection against production relies only on the classification in profiles.yml, with no additional explicit safeguard.
Static scan
- info risk
bundled scripts, not run by the SKILL.md
scripts/sandbox/classify_sandbox.py, scripts/sandbox/detect_hardcoded_db.py, scripts/sandbox/parse_profiles.py, scripts/sandbox/readonly_check.py, scripts/sandbox/substitute_placeholders.py
Original description
Shift-left safety net for dbt/SQL model edits. Runs change impact assessment before edits, generates SQL validation queries after, and executes them via `/mc-validate run`. Delegates health and monitor creation to peer skills.
added on Oct 19, 2026 · license Apache-2.0