io.github.machinavitalis/jaxonomy

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. jaxonomy/_init.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

  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 (jaxonomy/simulation/provenance.py: subprocess.* call; jaxonomy/dashboard/api.py: outbound HTTP/socket call). has both shell/exec and destructive filesystem writes (jaxonomy/simulation/provenance.py: subprocess.* call; jaxonomy/cli/run_fit_sindy.py: open(…, 'w'/'a'/'x') write-mode file access). discloses shell/exec capability: jaxonomy/simulation/provenance.py: subprocess.* call discloses dynamic code execution: 29 files (jaxonomy/analysis/block_jacobians.py, jaxonomy/backend/_jax/results_data.py, jaxonomy/backend/results_data.py, +26 more) discloses broad/destructive filesystem access: 9 files (jaxonomy/cli/run_fit_sindy.py, jaxonomy/cli/run_jaxonomy.py, jaxonomy/cli/run_variants.py, +6 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 3.4.0 (patch bump): new capability primitive(s) in source: child_process, dynamic-eval, fs-write, http-client, subprocess; new hardcoded outbound host/BCC destination(s): app.jaxonomy.com, arxiv.org, coin-or.github.io, doc.modelica.org, docs.kidger.site, docs.python.org, docs.scipy.org, doi.org, drake.mit.edu, en.wikipedia.org, github.com, jax.readthedocs.io, jaxonomy.atlassian.net, modelica.org, numpy.org, optax.readthedocs.io, py.jaxonomy.com, pypi.org, raw.githubusercontent.com, shorturl.at, specification.modelica.org, underactuated.csail.mit.edu, web.mit.edu, www.apache.org, www.engineeringtoolbox.com, www.firedrakeproject.org, www.iapws.org, www.jaxonomy.com, www.researchgate.net.

• 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 tracking behavior — disclosure