io.github.hashgraph-online/hol-guard

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

  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. ci/native_runtime/native_hook_client_support.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.

  references credential paths: dashboard/src/evidence/categories.tsx: references a credential path (id_rsa) references credential paths: rust/crates/guard-command/src/pretool.rs: references a credential path (/.ssh/) has both shell/exec and network egress — capable of download-and-run, though no such path was observed (ci/native_runtime/native_hook_client_support.py: subprocess.* call; ci/native_runtime/native_hook_client_support.py: outbound HTTP/socket call). has both shell/exec and destructive filesystem writes (ci/native_runtime/native_hook_client_support.py: subprocess.* call; ci/native_runtime/resident_test_support.py: destructive filesystem call (shutil.rmtree/os.remove/etc.)). discloses shell/exec capability: ci/native_runtime/native_hook_client_support.py: subprocess.* call discloses shell/exec capability: ci/native_runtime/probe_native_default_auto.py: subprocess.* call discloses shell/exec capability: distributions/mcpb/server/main.py: subprocess.* call discloses dynamic code execution: dashboard/src/apps/app-catalog.ts: uses eval(…) discloses broad/destructive filesystem access: ci/native_runtime/resident_test_support.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: dashboard/vite.config.ts: builds a path containing '..' (traversal-prone) 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 3.0.53 (patch bump): new capability primitive(s) in source: child_process, dynamic-eval, http-client, net, subprocess; new hardcoded outbound host/BCC destination(s): 10.0.0.1, 127.0.0.1, 127.1, 172.15.0.1, 172.16.0.1, 172.31.255.255, 192.168.1.1, 2130706433, api.example.com, api.example.invalid, attacker.example, cloud.example.test, evil.com, example.com, example.invalid, example.test, git-scm.com, github.com, hol.org, huggingface.co, private.example, twitter.com, upload.example.invalid, user.

• 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 telemetry behavior — disclosure