ai.dinglebear/unraid-mcp

What we checked

Six static checks, weighted by risk. Every result reflects only what could be observed in the published package and repository — never intent.

• Injection surface

  unscannable25% of grade

  Tool descriptions/manifest scanned for instruction-injection patterns (imperatives at the model, hidden text, 'ignore previous', data-exfil URLs).

  No tool descriptions, server instructions, prompts, or resources were fetched; nothing to scan for injection.

• Supply chain

  pass25% of grade

  Package provenance: namespace verification, repo linkage, maintainer count, account age, postinstall scripts, typosquat distance.

  Python build script has no build-time network/exec; repository linkage verified

• Credential hygiene

  warn15% of grade

  How the server takes secrets (env vs plaintext config vs hardcoded); secrets appearing in tool schemas.

  tests/test_auth.py: assigns a literal to token (corr…en (len 13)) — non-provider secret literal; prefer environment intake. tests/test_auth.py: assigns a literal to token (my-s…en (len 15)) — non-provider secret literal; prefer environment intake. tests/test_live.sh: assigns a literal to TOKEN (${UN…}} (len 38)) — non-provider secret literal; prefer environment intake. tests/test_live.sh: assigns a literal to TOKEN (ci-i…en (len 20)) — non-provider secret literal; prefer environment intake. tests/http_layer/test_request_construction.py: assigns a literal to API_KEY (test…45 (len 18)) — non-provider secret literal; prefer environment intake. tests/integration/test_subscriptions.py: assigns a literal to api_key (my-s…ey (len 13)) — non-provider secret literal; prefer environment intake. Reads secrets from the environment (e.g. scripts/generate_unraid_api_reference.py) — the recommended intake shape.

• Permission scope

  warn15% of grade

  Declared tools vs. breadth (filesystem, network, exec); flags shell-exec and unbounded filesystem access.

  has both shell/exec and network egress — capable of download-and-run, though no such path was observed (scripts/generate_unraid_api_reference.py: subprocess.* call; scripts/generate_unraid_api_reference.py: outbound HTTP/socket call). discloses shell/exec capability: scripts/generate_unraid_api_reference.py: subprocess.* call Capability DISCLOSURE, not a verdict: static analysis sees the primitive is present and reachable, not whether its use is attacker-controlled.

• Version behavior

  fail10% of grade

  Diff of tool definitions between versions; new permissions or changed descriptions in a patch release (the postmark-mcp class).

  vs prior 2.10.0 (patch bump): new capability primitive(s) in source: http-client, net, subprocess; new hardcoded outbound host/BCC destination(s): 10.0.0.1, 10.0.0.5, 10.1.0.2, 100.118.209.1, 127.0.0.1, 169.254.169.254, 192.168.1.1, 192.168.1.10, 192.168.1.100, account.unraid.net, accounts.google.com, evil.example, evil.example.com, example.com, from-dotenv, from-shell, github.com, gofastmcp.com, host, mcp.example.com, my-unraid.example.com, myserver.com, myserver.example.com, myunraid.net, new.url, no-such-host.invalid, oauth2.googleapis.com, ok.test, old.url, quoted.test, real.url, symlinked-sentinel, the-guild.dev, tower, tower.example.com, tower2.test, trusted.example.com, unraid-mcp.example.com, user, www.googleapis.com, x, your-unraid-server.

• Transport config

  warn10% of grade

  Remote servers: TLS and auth mode (none/token/OAuth). Local servers: whether the manifest indicates it phones home.

  manifest/source references tracking behavior — disclosure