io.github.yxy050208/multisim-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.

  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_cli.py: assigns a literal to secret (cli-…ak (len 24)) — non-provider secret literal; prefer environment intake. tests/test_provider_config.py: assigns a literal to secret (prob…ue (len 18)) — non-provider secret literal; prefer environment intake. tests/test_provider_config.py: assigns a literal to secret (secr…th (len 14)) — non-provider secret literal; prefer environment intake. Reads secrets from the environment (e.g. multisim_mcp/cli.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 (multisim_mcp/com_worker_client.py: subprocess.* call; multisim_mcp/com_worker.py: outbound HTTP/socket call). has both shell/exec and destructive filesystem writes (multisim_mcp/com_worker_client.py: subprocess.* call; multisim_mcp/agent_audit.py: open(…, 'w'/'a'/'x') write-mode file access). discloses shell/exec capability: 4 files (multisim_mcp/com_worker_client.py, multisim_mcp/job_engine.py, multisim_mcp/multisim_client.py, +1 more) discloses broad/destructive filesystem access: 8 files (multisim_mcp/agent_audit.py, multisim_mcp/design_patch_transactions.py, multisim_mcp/design_patch_workflow.py, +5 more) 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 1.1.0 (minor bump): new capability primitive(s) in source: fs-write, http-client, subprocess; new hardcoded outbound host/BCC destination(s): 127.0.0.1, api.deepseek.com, api.openai.com, attacker.invalid, example.invalid, github.com, models.example.test, other.example.test, www.w3.org.

• Transport config

  warn10% of grade

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

  sends telemetry to third-party SDK (amplitude) — disclosure, verify opt-in; manifest/source references heartbeat behavior — disclosure