io.github.0xzr/freellmpool

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. examples/agent_openai_sdk.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 (scripts/check_release_ready.py: subprocess.* call; integrations/opencode/freellmpool.js: outbound HTTP/socket call). has both shell/exec and destructive filesystem writes (scripts/check_release_ready.py: subprocess.* call; scripts/check_release_ready.py: destructive filesystem call (shutil.rmtree/os.remove/etc.)). discloses shell/exec capability: scripts/check_release_ready.py: subprocess.* call discloses broad/destructive filesystem access: scripts/check_release_ready.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: src/freellmpool/quota.py: open(…, 'w'/'a'/'x') write-mode file access discloses broad/destructive filesystem access: src/freellmpool/stats.py: open(…, 'w'/'a'/'x') write-mode file access 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 0.12.0 (minor bump): new capability primitive(s) in source: fs-write, http-client, subprocess; new hardcoded outbound host/BCC destination(s): 0.0.0.0, 017700000001, 0x7f000001, 0xzr.github.io, 10.0.0.5, 127, 127.0.0.1, 127.0.0.1., 127.1, 169.254.169.254, 192.168.1.1, 2130706433, 8.8.8.8, a.test, aider.chat, aistudio.google.com, alpha.test, api, api-inference.modelscope.cn, api.aionlabs.ai, api.cloudflare.com, api.evil.test, api.example.test, api.groq.com, api.hyperbolic.xyz, api.insecure.test, api.openai.com, api.test, api2.example.test, artificialanalysis.ai, b.test, beta.test, build.nvidia.com, builtin.test, cloud.cerebras.ai, cloud.sambanova.ai, cloud.siliconflow.cn, cohere.test, console.groq.com, console.mistral.ai, custom.test, dash.cloudflare.com, dashboard.cohere.com, datasets-server.huggingface.co, embed.test, evil, evil.artificialanalysis.ai.attacker.com, evil.example, example.test, free.test, gee.test, github.com, glama.ai, lmarena.ai, localhost., longcat.chat, mcp.so, mcpmarket.com, modelscope.cn, ok.test, ollama.com, opencode.ai, openrouter.ai, p, p.test, plugin.test, raw.githubusercontent.com, router.huggingface.co, smithery.ai, static.modelcontextprotocol.io, text.pollinations.ai, token.llm7.io, user, vercel.com, w.test, whisper.test, www.morphllm.com, www.pulsemcp.com, www.w3.org, www.youtube.com, x.test, z.ai, z.test.

• Transport config

  fail10% of grade

  Remote servers: TLS and auth mode (none/token/OAuth). Local servers: whether the manifest indicates it phones home.

  source performs runtime self-update / download-and-execute (bypasses version_behavior)