io.github.dcondrey/misterdev

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_gatekeeper.py: contains a hardcoded Slack token (xoxb…aa (len 23)) — in an example/sample file; verify it is not a real key. tests/test_gatekeeper.py: assigns a literal to password (${DB…D} (len 14)) — non-provider secret literal; prefer environment intake. tests/test_gatekeeper.py: assigns a literal to api_key (os.e…Y] (len 15)) — non-provider secret literal; prefer environment intake. tests/test_gatekeeper.py: assigns a literal to token (proc…EN (len 17)) — non-provider secret literal; prefer environment intake. tests/test_gatekeeper.py: assigns a literal to API_KEY (sk-d…ef (len 19)) — non-provider secret literal; prefer environment intake. tests/test_gatekeeper.py: assigns a literal to password (hunt…o1 (len 15)) — non-provider secret literal; prefer environment intake. tests/test_gatekeeper.py: assigns a literal to SECRET (tops…23 (len 17)) — non-provider secret literal; prefer environment intake. Reads secrets from the environment (e.g. misterdev/core/evolution/adapters.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 (misterdev/agent.py: subprocess.* call; misterdev/core/economics/free_models.py: outbound HTTP/socket call). has both shell/exec and destructive filesystem writes (misterdev/agent.py: subprocess.* call; misterdev/core/economics/llm_cache.py: destructive filesystem call (shutil.rmtree/os.remove/etc.)). discloses shell/exec capability: 17 files (misterdev/agent.py, misterdev/core/context/change_tracker.py, misterdev/core/context/lsp_swift.py, +14 more) discloses dynamic code execution: misterdev/core/execution/container.py: uses eval()/exec() discloses broad/destructive filesystem access: 4 files (misterdev/core/economics/llm_cache.py, misterdev/core/planning/sovereign.py, misterdev/tools/file_io.py, +1 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 0.5.0 (minor bump): new capability primitive(s) in source: child_process, fs-write, http-client, subprocess; new hardcoded outbound host/BCC destination(s): api.github.com, api.npmjs.org, cdnjs.cloudflare.com, example, example.com, get.example.com, github.com, gw.example, openrouter.ai, pypi.org, raw.githubusercontent.com, registry.modelcontextprotocol.io, registry.npmjs.org, x.

• 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 (posthog) — disclosure, verify opt-in; manifest/source references tracking behavior — disclosure