io.github.trueoriginlabs/vibatchium

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_skills.py: embeds a PRIVATE KEY block. (in an example/sample file — lower severity, but a real key here still leaks.) Reads secrets from the environment (e.g. tests/conftest.py) — the recommended intake shape.

• Permission scope

  fail15% of grade

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

  HIGH: executes content derived from a network response (vibatchium/cli.py: passes a network response into an exec/eval call). references credential paths: vibatchium/setup_cmd.py: references a credential path (/.cursor/) has both shell/exec and destructive filesystem writes (vibatchium/cli.py: subprocess.* call; vibatchium/geo.py: destructive filesystem call (shutil.rmtree/os.remove/etc.)). discloses shell/exec capability: 7 files (vibatchium/cli.py, vibatchium/cli.py, vibatchium/client.py, +4 more) discloses broad/destructive filesystem access: 14 files (vibatchium/geo.py, vibatchium/gpu.py, vibatchium/login.py, +11 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 0.19.1 (patch bump): new capability primitive(s) in source: child_process, fs-write, http-client, net, subprocess; new hardcoded outbound host/BCC destination(s): ..., 10.0.0.1, 127.0.0.1, 127.0.0.1.evil.com, 127.0.0.5, 169.254.169.254, 8.8.8.8, abrahamjuliot.github.io, ali, anything.com, api.ipify.org, bot.sannysoft.com, brd.superproxy.io, docs.example.com, docs.python.org, does-not-exist-zzz123456.invalid, e.com, evil.test, ex.com, example.com, gate.decodo.com, geminixprize.com, geo.iproyal.com, github.com, h, hackerone.com, hooks.example.com, host, kaliiiiiiiiii-vinyzu.github.io, localhost.evil.com, m.youtube.com, maybe-dead.example, music.youtube.com, nope.cloudflare-site.test, nowhere.example, secretuser, sub.example.com.au, this-host-definitely-does-not-exist-xyz.invalid, this-host-definitely-does-not-exist-xyz123.invalid, u, user, www.aliexpress.com, www.amazon.com, www.ex.com, www.example.co.uk, www.example.com, www.github.com, x, x.com, youtube.com.

• 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 telemetry behavior — disclosure