← All scanned servers

Trust Card · Static snapshot

mcp-memory D

54/100 · 5 findings (0 critical) · scope broad · Ranked #38 of 53 scanned.

Scanned 2026-09-09 · engine v0.1.0 (8 regex rules, OWASP-mapped) · upstream modelcontextprotocol/servers @ d73f99e (2026-09-02)

Score breakdown

security 25
permissions 60
provenance 65
reliability 85
stability 75

Score constrained by 0 critical and 5 high-severity findings.

Fix first (5 critical/high)

SeverityRuleFindingRemediation
HIGHAT-SEC-005 (ASI02)Unrestricted Recursive File Deletion / Modification
__tests__/delete-reporting.test.ts:35
Evidence: await fs.unlink(testFilePath);
Enforce strict jail/root directories and require explicit human-in-the-loop confirmation before file deletions.
HIGHAT-SEC-005 (ASI02)Unrestricted Recursive File Deletion / Modification
__tests__/file-path.test.ts:32
Evidence: await fs.unlink(oldMemoryPath);
Enforce strict jail/root directories and require explicit human-in-the-loop confirmation before file deletions.
HIGHAT-SEC-005 (ASI02)Unrestricted Recursive File Deletion / Modification
__tests__/file-path.test.ts:37
Evidence: await fs.unlink(newMemoryPath);
Enforce strict jail/root directories and require explicit human-in-the-loop confirmation before file deletions.
HIGHAT-SEC-005 (ASI02)Unrestricted Recursive File Deletion / Modification
__tests__/knowledge-graph.test.ts:23
Evidence: await fs.unlink(testFilePath);
Enforce strict jail/root directories and require explicit human-in-the-loop confirmation before file deletions.
HIGHAT-SEC-005 (ASI02)Unrestricted Recursive File Deletion / Modification
index.ts:197
Evidence: await fs.unlink(tempFilePath).catch(() => {});
Enforce strict jail/root directories and require explicit human-in-the-loop confirmation before file deletions.

Permissions

Scope: broad · Shell: disabled · Network egress: no · File deletion: enabled · Human approval: none

Provenance

License: SEE LICENSE IN LICENSE · Lockfile: no · Security policy: no · Signals: unverified origin

Independently scanned by the AgentTrust registry (not self-reported by the project). Static analysis only — no code executed, findings need human triage, counts may include test/example code. Static snapshot; re-scan before relying on it: npx @eulogik/agenttrust scan https://github.com/modelcontextprotocol/servers/tree/main/src/memory. Scores move with every upstream commit; pages refresh weekly. How scoring works.