io.github.RudrenduPaul/agent-observability

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: subprocess in build script, 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.

  examples/04-google-genai-thinking-config/example.py: assigns a literal to api_key (demo…ed (len 17)) — non-provider secret literal; prefer environment intake. examples/16-openai-sdk-plain-embeddings/example.py: assigns a literal to api_key (fake…ed (len 17)) — non-provider secret literal; prefer environment intake. examples/17-langgraph-toolnode-custom-provider/example.py: assigns a literal to api_key (fake…ed (len 17)) — non-provider secret literal; prefer environment intake. tests/integration/test_google_genai.py: assigns a literal to api_key (fake…ed (len 19)) — non-provider secret literal; prefer environment intake. tests/unit/test_botocore_hook.py: contains a hardcoded AWS access key id (AKIA…KE (len 20)) — in an example/sample file; verify it is not a real key. tests/unit/test_botocore_hook.py: contains a hardcoded AWS secret access key (aws_…e" (len 64)) — in an example/sample file; verify it is not a real key. tests/unit/test_replay_engine.py: contains a hardcoded AWS secret access key (aws_…e" (len 64)) — in an example/sample file; verify it is not a real key. Reads secrets from the environment (e.g. examples/02-langgraph-failure-replay/example.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 (npm/bin/agent-trace.js: imports child_process and calls exec/spawn; examples/02-langgraph-failure-replay/example.py: outbound HTTP/socket call). discloses shell/exec capability: npm/bin/agent-trace.js: imports child_process and calls exec/spawn discloses shell/exec capability: src/agent_trace/_cli.py: subprocess.* call discloses shell/exec capability: src/agent_trace/mcp_server.py: subprocess.* call discloses dynamic code execution: examples/04-agno-agent-team/example.py: uses eval()/exec() 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.RudrenduPaul/[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

  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