Prefect (PyPI) MCP Server
mcp-prefect
PyPI
v0.2.3419
Published by an unidentified publisher — no publish provenance and no public repository, so the publisher could not be verified and the source cannot be independently located.
MCP Server for Prefect
The grade answers one question — how safe is this server for you to adopt — so it is computed in two auditable stages. Nothing below is an opinion or an LLM's guess; every line is a real term the deterministic engine applied, and the same input always yields the same number.
1. Threat score — 100 − 0 = 100. What the published surface and source actually contain:
The deterministic scan raised no scored threat in the surface it inspected — the threat score stayed at 100. Capability observations and advisory notes are recorded but never lower it.
2. Client adoption risk — 100 − 8 = 92. Three small, subtract-only factors that reflect your risk in adopting it — a clean scan proves less on a powerful, unverified or barely-inspectable package, so the grade says so plainly:
| Points | Adoption-risk factor |
|---|---|
| −6 | capability blast radius (high) — client exposure if the model is manipulated |
| −2 | publisher verification (unlinked) — no provenance/repo link, but the shipped source was fully read |
Capability observations and info notes are shown under Findings but never scored.
Open any row's finding below for the file, line and evidence behind a deduction.
Findings 4
high Dynamic evaluation of a non-literal value (test_scripts/test_more_prefect.py)MTC-SRC-010
In the server's implementation (`test_scripts/test_more_prefect.py:19`): Evaluating a runtime value as code (rather than a fixed literal) executes whatever reaches it — a direct RCE primitive, and almost never necessary in legitimate code. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: xt", "") data = eval(text) return data.get("state", {}).get("type", "UNKNOWN") return "UNKNOWN" as
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server test_scripts/test_more_prefect.py
high Dynamic evaluation of a non-literal value (test_scripts/test_prefect.py)MTC-SRC-010
In the server's implementation (`test_scripts/test_prefect.py:142`): Evaluating a runtime value as code (rather than a fixed literal) executes whatever reaches it — a direct RCE primitive, and almost never necessary in legitimate code. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: flows_data = eval(content.text) if flows_data["flows"]: f
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server test_scripts/test_prefect.py
high Dynamic code execution in server code (test_scripts/test_more_prefect.py)MTC-SRC-001
In the server's implementation (`test_scripts/test_more_prefect.py:19`): Evaluating strings as code is the most direct RCE primitive; if any tool input reaches it, the server executes attacker-chosen code. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: xt", "") data = eval(text) return data.get("state", {}).get("type", "UNKNOWN") return "UNKNOWN" as
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server test_scripts/test_more_prefect.py
high Dynamic code execution in server code (test_scripts/test_prefect.py)MTC-SRC-001
In the server's implementation (`test_scripts/test_prefect.py:142`): Evaluating strings as code is the most direct RCE primitive; if any tool input reaches it, the server executes attacker-chosen code. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: flows_data = eval(content.text) if flows_data["flows"]: f
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server test_scripts/test_prefect.py
Tools 46
Each tool and what it can reach — statically extracted from the published source.
cancel_flow_runno sensitive capabilitycreate_flow_run_from_deploymentno sensitive capabilitycreate_variableno sensitive capabilitycreate_work_queueno sensitive capabilitydelete_block_documentno sensitive capabilitydelete_deploymentno sensitive capabilitydelete_flowno sensitive capabilitydelete_flow_runno sensitive capabilitydelete_variableno sensitive capabilitydelete_work_queueno sensitive capability
Show 36 more tools ↓
get_block_documentno sensitive capabilityget_block_documentsno sensitive capabilityget_block_typeno sensitive capabilityget_block_typesno sensitive capabilityget_current_workspaceno sensitive capabilityget_deploymentno sensitive capabilityget_deployment_scheduleno sensitive capabilityget_deploymentsno sensitive capabilityget_flowno sensitive capabilityget_flow_runno sensitive capabilityget_flow_runsno sensitive capabilityget_flow_runs_by_flowno sensitive capabilityget_flowsno sensitive capabilityget_healthno sensitive capabilityget_task_runno sensitive capabilityget_task_runsno sensitive capabilityget_task_runs_by_flow_runno sensitive capabilityget_variableno sensitive capabilityget_variablesno sensitive capabilityget_work_queueno sensitive capabilityget_work_queue_by_nameno sensitive capabilityget_work_queuesno sensitive capabilityget_workspaceno sensitive capabilityget_workspace_by_handleno sensitive capabilityget_workspacesno sensitive capabilitypause_deployment_scheduleno sensitive capabilitypause_work_queueno sensitive capabilityrestart_flow_runno sensitive capabilityresume_deployment_scheduleno sensitive capabilityresume_work_queueno sensitive capabilityset_deployment_scheduleno sensitive capabilityset_flow_run_stateno sensitive capabilityset_task_run_stateno sensitive capabilityupdate_deploymentno sensitive capabilityupdate_variableno sensitive capabilityupdate_work_queueno sensitive capability
What this scan could not see
- Tools were statically extracted from the published source (46 recovered), not enumerated from a running server. Tool-poisoning, Unicode-smuggling, capability and toxic-flow analysis ran on this inferred surface, but a mis-parsed registration could be missed or mis-attributed, so tool-derived findings are capped below “confirmed”. To grade the real runtime surface, scan the running server: --command "npx -y <package>".
Versions 1
Scan history per published version. The engine is deterministic — the same version always yields the same score, so a changed score means the package itself changed.
| Version | Score | Findings | Engine | Scanned |
|---|---|---|---|---|
v0.2.3419 latest |
A 92/100 | 4 | 1.13.0 | 2026-08-25 |
Embed this score
Show this server's live Trust Score in your README, docs or website. The badge is served straight from the registry and updates automatically after every rescan — no API key needed. It links back to this page, so anyone who sees the grade can also read the findings behind it instead of taking a number on faith.
Verify this score yourself
The score above is reproducible: the same package version always yields the same result. Run it locally or over the free API — no account, no LLM, fully deterministic.
npx mcptrustchecker scan mcp-prefect --online --registry pypiOther implementations of Prefect 3
Independent packages implementing the same tool, scanned with the same engine. Compare all 4 side by side →
MCP server for interacting with the Prefect API
MCP server for interacting with Prefect
Security scan results for the Prefect (npm) MCP server.
More in Other
A Model Context Protocol server implementation for AdGuard Home that enables AI agents to query and manage DNS records, filtering rules and more
Identity oracle and trust layer for autonomous AI agents. Bidirectional KYA and trust scoring.
MCP server with Airtable integration
A free one-minute reset for people, delivered by AI, with no account or personal data.
Access SEC filings efficiently, save time and tokens, and get cited answers.
A smart [MCP](https://modelcontextprotocol.io) server for [AniList](https://anilist.co) that gets your anime/manga taste - not just API calls.