Unreleased: not public yet. Maintainers have not reviewed this result.
Results / @modelcontextprotocol/server-postgres
@modelcontextprotocol/server-postgres
A
Protocol4/4 points
D
Database4/7 pointscritical: mutating-sql
Coverage: covered (1 of 1 tools reached with a valid call).
Protocol checks
| Check | Verdict | What happened |
|---|---|---|
| protocol Valid MCP handshake and tool schemas | pass | handshake ok, 1 well-formed tool(s) |
| tool-poisoningcritical No tool-poisoning indicators in tool metadata | pass | scanned 1 tool definition(s), 7 resource(s), including parameter text: nothing suspicious |
| robust-input Survives malformed tool input | pass | 3 invalid call(s) rejected cleanly |
| secret-echocritical No environment secret echoed in responses | pass | canary secrets absent from tools/list and 3 captured response(s) |
Database checks
| Check | Verdict | What happened |
|---|---|---|
| read-only Read-only by default | pass | plain writes blocked on all 1 SQL tool(s) |
| write-gate (advisory) Write approval gate | n/a | no write concept: no writer tool declared and every reader blocked writes |
| rlscritical Row-level security respected | pass | DB-enforced: the server passed the policy through and added no filtering of its own (support role saw only its own tenant on every tool) |
| mutating-sqlcritical Mutating SQL blocked | fail | EXECUTED: COMMIT then write (read-only transaction escape); 15 of 16 blocked |
| piicritical Personal data masking respected | pass | DB-enforced: the server passed the grants through and added no filtering of its own (analyst role never received raw emails on any tool) |
| pii-masked Masked data reachable | pass | the masked view was readable and returned only masked emails |
| db-rowcap Row cap | fail | no row cap: 200,000 rows could be read through every tool |
| db-timeout Statement timeout | fail | no statement timeout within 18s (expected 10s): SELECT pg_sleep(30) on query |
| audit (advisory) Every call logged | n/a | not measured: servers do not write audit rows by themselves |
| db-privilege (advisory) Connects as the database role it was given | pass | every instance connected as exactly the role it was given (sm_admin, sm_support, sm_analyst); none is a superuser, has BYPASSRLS or owns an RLS table |
What was not covered
- HTTP: Not applicable. The server is started over stdio; the http pack needs server.url
- Files: Not applicable. No tool is classified as taking a file path with medium confidence or higher
- Git: Not applicable. No tool is classified as taking a git repository path with medium confidence or higher
- Network: Not applicable. No tool is classified as taking a URL with medium confidence or higher
- SQLite: Not applicable. No {sqlite_path} placeholder in the server config, so the harness cannot give the server a database to grade (add {sqlite_path} to the server args)
- Effects: Not applicable. No tool takes a harness-owned target (file path, repository or URL) with medium confidence or higher
- Tools that a server lists only for clients that declare roots, sampling or elicitation capabilities are not seen: the harness declares none
- The sandbox is a convention the harness follows, not a security boundary. A malicious server runs with your user's privileges, sees the real file system and network, and can ignore cwd, HOME and PATH. Run untrusted servers in a throwaway VM or container.
- Prompt injection carried in tool output (only tool metadata is scanned)
- Authentication and transport security
- Audit logging (not measured by these packs; only the db pack has an advisory audit check)
Run it yourself
Started with the standard settings like this; put it in sealmark.yaml as the server:
npx -y @modelcontextprotocol/server-postgres@0.6.2
Each of 613 verdicts was checked independently, and Sealmark was wrong on some: how reliable Sealmark is.
Think a verdict is wrong, or want a server removed? Email disclosure@sealmark.dev with the server version. We recheck, correct the page if we were wrong, and add your reply. See the disclosure policy.