io.github.iowarp/jarvis-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

  pass15% of grade

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

  Reads secrets from the environment (e.g. clio-kit-mcp-servers/adios/src/adios_mcp/server.py) — the recommended intake shape. PASS = no static red flag; static analysis cannot prove the code honors env-based secret handling at runtime.

• Permission scope

  fail15% of grade

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

  HIGH: downloads and pipes straight to a shell (clio-kit-mcp-servers/paraview/scripts/installation_example.py: pipes a download into a shell). has both shell/exec and destructive filesystem writes (clio-kit-mcp-servers/chronolog/src/chronomcp/utils/helpers.py: subprocess.* call; clio-kit-mcp-servers/arxiv/src/arxiv_mcp/capabilities/download_paper.py: open(…, 'w'/'a'/'x') write-mode file access). discloses shell/exec capability: 20 files (clio-kit-mcp-servers/chronolog/src/chronomcp/utils/helpers.py, clio-kit-mcp-servers/node-hardware/src/node_hardware_mcp/capabilities/remote_node_info.py, clio-kit-mcp-servers/node-hardware/src/node_hardware_mcp/capabilities/utils.py, +17 more) discloses dynamic code execution: clio-kit-mcp-servers/hdf5/src/hdf5_mcp/server.py: uses eval()/exec() discloses broad/destructive filesystem access: 16 files (clio-kit-mcp-servers/arxiv/src/arxiv_mcp/capabilities/download_paper.py, clio-kit-mcp-servers/chronolog/src/chronomcp/capabilities/retrieve_handler.py, clio-kit-mcp-servers/compression/src/compression_mcp/capabilities/compression_base.py, +13 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 3.6.1 (patch bump): new capability primitive(s) in source: child_process, dynamic-eval, fs-write, http-client, subprocess; new hardcoded outbound host/BCC destination(s): 155.101.6.191, adios2.readthedocs.io, api.example.com, arxiv.org, astral.sh, code.visualstudio.com, cursor.com, docs.hdfgroup.org, example.com, export.arxiv.org, github.com, gitlab.kitware.com, grc.iit.edu, iowarp.ai, matplotlib.org, new.nsf.gov, pandas.pydata.org, parquet.apache.org, slurm.schedmd.com, test.example.com, toolkit.iowarp.ai, www.claude.com, www.hdfgroup.org, www.mcs.anl.gov, 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.

  manifest/source references analytics behavior — disclosure