io.github.Cohexa-ai/stale-write-guard-fs

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.

  src/ccs/adapters/substrate.py: assigns a literal to token (Subs…en (len 14)) — non-provider secret literal; prefer environment intake. tests/test_remote_transport_guard.py: assigns a literal to secret (tops…ue (len 22)) — non-provider secret literal; prefer environment intake. Reads secrets from the environment (e.g. src/ccs/adapters/claude_code/auth.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: shell/exec capability AND network egress together — download-and-execute risk (src/ccs/adapters/claude_code/resolver.py: subprocess.* call; src/ccs/adapters/claude_code/lifecycle.py: outbound HTTP/socket call). HIGH: shell/exec capability AND destructive filesystem writes together (src/ccs/adapters/claude_code/resolver.py: subprocess.* call; src/ccs/adapters/claude_code/coordinator_server.py: path traversal ('../') combined with file access). discloses shell/exec capability: src/ccs/adapters/claude_code/resolver.py: subprocess.* call discloses shell/exec capability: src/ccs/cli/coherence_coordinator.py: subprocess.* call discloses shell/exec capability: src/ccs/hardening/release_readiness.py: subprocess.* call discloses shell/exec capability: tests/test_claude_code_cli.py: subprocess.* call discloses shell/exec capability: tests/test_claude_code_e2e.py: subprocess.* call discloses shell/exec capability: tests/test_claude_code_resolver.py: subprocess.* call discloses shell/exec capability: tests/test_coherence_client_tls.py: subprocess.* call discloses shell/exec capability: tests/test_demo_script.py: subprocess.* call discloses shell/exec capability: tests/test_diagnose_calibration.py: subprocess.* call discloses shell/exec capability: tests/test_diagnose_cli.py: subprocess.* call discloses shell/exec capability: tests/test_diagnose_no_import_side_effects.py: subprocess.* call discloses shell/exec capability: tests/test_single_file_demos.py: subprocess.* call discloses shell/exec capability: pyproject declares an exec-oriented dependency (pexpect/fabric/invoke) discloses dynamic code execution: tests/test_claude_code_lifecycle.py: uses eval/exec/__import__ discloses dynamic code execution: tests/test_state_transitions_log.py: uses eval/exec/__import__ discloses broad/destructive filesystem access: src/ccs/adapters/claude_code/coordinator_server.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: src/ccs/adapters/coherent_volume.py: open(…, 'w'/'a'/'x') write-mode file access discloses broad/destructive filesystem access: src/ccs/adapters/coherent_volume.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: src/ccs/cli/coherence_hook_client.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: src/ccs/cli/coherence_replay.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: src/ccs/coordinator/backend_contract.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: src/ccs/coordinator/sqlite_registry.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: src/ccs/diagnose/telemetry.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: src/ccs/replay/recorder.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: tests/test_claude_code_coordinator_server.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: tests/test_claude_code_lifecycle.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: tests/test_claude_code_policy.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: tests/test_claude_code_policy_strict_mode.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: tests/test_coherence_client_tls.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: tests/test_diagnose_calibration.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: tests/test_refactor_demo.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: tests/test_replay_resolver.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: tests/test_sqlite_registry.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) 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.Cohexa-ai/[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 telemetry behavior — disclosure