io.github.Mibayy/token-savior-recall

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_config_analyzer.py: embeds a PRIVATE KEY block. (in an example/sample file — lower severity, but a real key here still leaks.) Reads secrets from the environment (e.g. hooks/bash_rewriter_hook.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: src/token_savior/cli_init/agent_paths.py: references a sensitive credential path (/.cursor/) — the tool-poisoning exfiltration target. HIGH: tests/test_cli_init.py: references a sensitive credential path (/.cursor/) — the tool-poisoning exfiltration target. HIGH: shell/exec capability AND network egress together — download-and-execute risk (benchmarks/run_benchmarks.py: subprocess.* call; src/token_savior/cli.py: outbound HTTP/socket call). HIGH: shell/exec capability AND destructive filesystem writes together (benchmarks/run_benchmarks.py: subprocess.* call; benchmarks/run_benchmarks.py: open(…, 'w'/'a'/'x') write-mode file access). discloses shell/exec capability: benchmarks/run_benchmarks.py: subprocess.* call discloses shell/exec capability: hooks/openclaw/token-savior-memory/handler.js: imports child_process and calls exec/spawn discloses shell/exec capability: scripts/export_markdown.py: subprocess.* call discloses shell/exec capability: src/token_savior/breaking_changes.py: subprocess.* call discloses shell/exec capability: src/token_savior/cli.py: subprocess.* call discloses shell/exec capability: src/token_savior/git_tracker.py: subprocess.* call discloses shell/exec capability: src/token_savior/impacted_tests.py: subprocess.* call discloses shell/exec capability: src/token_savior/project_actions.py: subprocess.* call discloses shell/exec capability: src/token_savior/memory/consistency.py: subprocess.* call discloses shell/exec capability: src/token_savior/memory/turn_capture.py: subprocess.* call discloses shell/exec capability: src/token_savior/server_handlers/memory.py: subprocess.* call discloses shell/exec capability: tests/test_breaking_changes.py: subprocess.* call discloses shell/exec capability: tests/test_incremental.py: subprocess.* call discloses shell/exec capability: tests/test_slot_manager.py: subprocess.* call discloses dynamic code execution: hooks/memory-session-start.sh: uses eval/exec/__import__ discloses dynamic code execution: tests/test_memory_vector_setup.py: uses eval/exec/__import__ discloses broad/destructive filesystem access: benchmarks/run_benchmarks.py: open(…, 'w'/'a'/'x') write-mode file access discloses broad/destructive filesystem access: benchmarks/run_benchmarks.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: hooks/openclaw/token-savior-memory/handler.js: destructive/writing fs.* call (write/unlink/rm/chmod/rename) discloses broad/destructive filesystem access: hooks/openclaw/token-savior-memory/handler.js: path traversal ('../') combined with file access discloses broad/destructive filesystem access: src/token_savior/cache_ops.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: src/token_savior/checkpoint_ops.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: src/token_savior/checkpoint_ops.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: src/token_savior/cli.py: open(…, 'w'/'a'/'x') write-mode file access discloses broad/destructive filesystem access: src/token_savior/cli.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: src/token_savior/server_runtime.py: open(…, 'w'/'a'/'x') write-mode file access discloses broad/destructive filesystem access: src/token_savior/server_handlers/memory.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: tests/test_breaking_changes.py: open(…, 'w'/'a'/'x') write-mode file access discloses broad/destructive filesystem access: tests/test_breaking_changes.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: tests/test_cache_ops.py: open(…, 'w'/'a'/'x') write-mode file access discloses broad/destructive filesystem access: tests/test_cache_ops.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: tests/test_cli_init.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: tests/test_daemon_client.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: tests/test_markup_ruby.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: tests/test_openclaw_hook_pack.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: tests/test_project_indexer.py: open(…, 'w'/'a'/'x') write-mode file access discloses broad/destructive filesystem access: tests/test_server_integration.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: tests/test_watcher.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.Mibayy/[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