io.github.Aitherium/aither-adk

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

  warn15% of grade

  How the server takes secrets (env vs plaintext config vs hardcoded); secrets appearing in tool schemas.

  adk/elysium.py: assigns a literal to api_key (aith…xx (len 18)) — non-provider secret literal; prefer environment intake. adk/faculties/base.py: assigns a literal to secret (aith…lt (len 26)) — non-provider secret literal; prefer environment intake. adk/faculties/code_graph.py: assigns a literal to secret (aith…lt (len 26)) — non-provider secret literal; prefer environment intake. Reads secrets from the environment (e.g. install.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: adk/mcp_setup.py reads a credential path, executes commands, and makes network calls in the same file — the combination that makes exfiltration possible. HIGH: downloads and pipes straight to a shell (setup-vllm.py: pipes a download into a shell). references credential paths: adk/mcp_setup.py: references a credential path (/mcp.json) references credential paths: adk/shell/mcp_setup.py: references a credential path (/mcp.json) installer script pipes a download into a shell — conventional, but worth reading before running (install.sh: pipes a download into a shell). has both shell/exec and destructive filesystem writes (install.py: subprocess.* call; adk/admin_api.py: destructive filesystem call (shutil.rmtree/os.remove/etc.)). discloses shell/exec capability: 61 files (install.py, setup-vllm.py, adk/addon_manager.py, +58 more) discloses dynamic code execution: 6 files (adk/builtin_tools.py, adk/agentdoc/_discover.py, adk/agentdoc/_structured.py, +3 more) discloses broad/destructive filesystem access: 44 files (adk/admin_api.py, adk/agent_daemon.py, adk/bugreport.py, +41 more) Capability DISCLOSURE, not a verdict: static analysis sees the primitive is present and reachable, not whether its use is attacker-controlled.

• Version behavior

  unscannable10% of grade

  Diff of tool definitions between versions; new permissions or changed descriptions in a patch release (the postmark-mcp class).

  no prior version to diff (first sight of io.github.Aitherium/[email protected]); version-drift / rug-pull cannot be evaluated. Cold-start risk is covered by point-in-time checks (injection_surface, supply_chain, credential_hygiene), not here.

• 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); manifest/source references telemetry behavior — disclosure