io.github.n24q02m/wet-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

  fail25% of grade

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

  Python build script executes at build time: network call in build script

• 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. src/wet_mcp/__main__.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 (src/wet_mcp/searxng_runner.py: subprocess.* call; src/wet_mcp/searxng_runner.py: outbound HTTP/socket call). has both shell/exec and destructive filesystem writes (src/wet_mcp/searxng_runner.py: subprocess.* call; src/wet_mcp/__main__.py: destructive filesystem call (shutil.rmtree/os.remove/etc.)). discloses shell/exec capability: src/wet_mcp/searxng_runner.py: subprocess.* call discloses shell/exec capability: src/wet_mcp/setup.py: subprocess.* call discloses shell/exec capability: src/wet_mcp/sync.py: subprocess.* call discloses broad/destructive filesystem access: src/wet_mcp/__main__.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: src/wet_mcp/sources/crawler.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) 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 3.10.0 (minor bump): new capability primitive(s) in source: http-client, net, subprocess; new hardcoded outbound host/BCC destination(s): 10.0.0.20, 127.0.0.1, angular.dev, api.github.com, api.nuget.org, apisix.apache.org, clerk.com, cmake.org, code.launchpad.net, crates.io, defold.com, developer.android.com, dhall-lang.org, docs., docs.celeryq.dev, docs.flutter.dev, docs.gitlab.com, docs.nestjs.com, docs.rs, docs.spline.design, docs.spring.io, echo.labstack.com, enhance.dev, fluxcd.io, github.com, graphql.org, grpc.io, guides.rubyonrails.org, hex.pm, hexdocs.pm, htmx.org, javadoc.io, kafka.apache.org, laravel.com, linkerd.io, nextjs.org, nginx.org, numpy.org, packagist.org, pandas.pydata.org, pinia.vuejs.org, pkg.go.dev, projectreactor.io, protobuf.dev, pub.dev, pypi.org, pytorch.org, raw.githubusercontent.com, redis.io, registry.npmjs.org, repo.packagist.org, router.vuejs.org, rubygems.org, scikit-learn.org, search.maven.org, sqitch.org, staticcheck.dev, supabase.com, symfony.com, tanstack.com, tekton.dev, trpc.io, turbo.build, vuejs.org, www.boost.org, www.chakra-ui.com, www.crummy.com, www.krakend.io, www.swift.org, www.tensorflow.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 tracking behavior — disclosure