io.github.agentic-hil/agentic-hil

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_redact.py: assigns a literal to token (alph…ta (len 13)) — non-provider secret literal; prefer environment intake. tests/test_test_reactor.py: assigns a literal to secret (oper…ak (len 29)) — non-provider secret literal; prefer environment intake. Reads secrets from the environment (e.g. src/agentic_hil/backends/common.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: downloads and pipes straight to a shell (src/agentic_hil/knowledge.py: pipes a download into a shell). has both shell/exec and destructive filesystem writes (src/agentic_hil/can.py: subprocess.* call; src/agentic_hil/backends/gdbdebug.py: destructive filesystem call (shutil.rmtree/os.remove/etc.)). discloses shell/exec capability: src/agentic_hil/can.py: subprocess.* call discloses shell/exec capability: src/agentic_hil/process.py: subprocess.* call discloses shell/exec capability: src/agentic_hil/upgrade.py: subprocess.* call discloses broad/destructive filesystem access: 5 files (src/agentic_hil/backends/gdbdebug.py, src/agentic_hil/backends/pyocd.py, src/agentic_hil/backends/stlink.py, +2 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 0.21.1 (patch bump): new capability primitive(s) in source: child_process, fs-write, http-client, net, subprocess; new hardcoded outbound host/BCC destination(s): agentic-hil.github.io, agentic-hil.readthedocs.io, buildbot, ci-bot, code.claude.com, example.invalid, github.com, h, opencode.ai, packages.example.internal, pyocd.io, pypi.example, pypi.org, 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 tracking behavior — disclosure