io.github.pretorin-ai/pretorin

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.

  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_agent_runner_coverage.py: assigns a literal to api_key (ptn-…ey (len 16)) — non-provider secret literal; prefer environment intake. tests/test_codex_agent.py: assigns a literal to api_key (ptn-…ey (len 16)) — non-provider secret literal; prefer environment intake. tests/evidence/test_redact.py: embeds a PRIVATE KEY block. (in an example/sample file — lower severity, but a real key here still leaks.) tests/evidence/test_redact.py: assigns a literal to password (supe…23 (len 14)) — non-provider secret literal; prefer environment intake. tests/evidence/test_redact.py: assigns a literal to API_KEY (abcd…34 (len 16)) — non-provider secret literal; prefer environment intake. Reads secrets from the environment (e.g. src/pretorin/attestation.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/pretorin/agent/codex_runtime.py: references a sensitive credential path (/mcp.json) — the tool-poisoning exfiltration target. HIGH: src/pretorin/agent/mcp_config.py: references a sensitive credential path (/mcp.json) — the tool-poisoning exfiltration target. HIGH: tools/cross-harness-smoke.py: references a sensitive credential path (/.cursor/) — the tool-poisoning exfiltration target. HIGH: shell/exec capability AND network egress together — download-and-execute risk (src/pretorin/attestation.py: subprocess.* call; docs/book/book-a0b12cfe.js: outbound HTTP/socket call). HIGH: shell/exec capability AND destructive filesystem writes together (src/pretorin/attestation.py: subprocess.* call; src/pretorin/plans.py: destructive filesystem call (shutil.rmtree/os.remove/etc.)). discloses shell/exec capability: src/pretorin/attestation.py: subprocess.* call discloses shell/exec capability: src/pretorin/preflight_discovery.py: subprocess.* call discloses shell/exec capability: src/pretorin/agent/codex_runtime.py: subprocess.* call discloses shell/exec capability: src/pretorin/cli/harness.py: subprocess.* call discloses shell/exec capability: src/pretorin/cli/main.py: subprocess.* call discloses shell/exec capability: src/pretorin/cli/recipe.py: subprocess.* call discloses shell/exec capability: tests/test_main_entry.py: subprocess.* call discloses shell/exec capability: tests/test_release_tools.py: subprocess.* call discloses shell/exec capability: tools/cross-harness-smoke.py: subprocess.* call discloses shell/exec capability: tools/smoke-binary.sh: subprocess.* call discloses dynamic code execution: src/pretorin/cli/artifacts.py: uses eval/exec/__import__ discloses dynamic code execution: src/pretorin/cli/context.py: uses eval/exec/__import__ discloses dynamic code execution: tests/test_main_entry.py: uses eval/exec/__import__ discloses dynamic code execution: tests/recipes/test_asset_inventory_recipes.py: uses eval/exec/__import__ discloses dynamic code execution: tests/recipes/test_ci_evidence_capture.py: uses eval/exec/__import__ discloses dynamic code execution: tools/render_homebrew_formula.py: uses eval/exec/__import__ discloses broad/destructive filesystem access: src/pretorin/plans.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: src/pretorin/plans.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: src/pretorin/preflight.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: src/pretorin/agent/codex_runtime.py: open(…, 'w'/'a'/'x') write-mode file access discloses broad/destructive filesystem access: src/pretorin/agent/mcp_config.py: open(…, 'w'/'a'/'x') write-mode file access discloses broad/destructive filesystem access: src/pretorin/cli/context.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: src/pretorin/cli/evidence.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: src/pretorin/cli/issues.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: src/pretorin/cli/monitoring.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: src/pretorin/cli/narrative.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: src/pretorin/cli/notes.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: src/pretorin/cli/review.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: src/pretorin/cli/skill.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: src/pretorin/cli/version_check.py: open(…, 'w'/'a'/'x') write-mode file access discloses broad/destructive filesystem access: src/pretorin/client/api.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: src/pretorin/client/config.py: open(…, 'w'/'a'/'x') write-mode file access discloses broad/destructive filesystem access: src/pretorin/mcp/handlers/stig.py: open(…, 'w'/'a'/'x') write-mode file access discloses broad/destructive filesystem access: src/pretorin/recipes/_recipes/openscap-baseline/scripts/run_scan.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: tests/test_api_client_plan_metadata.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: tests/test_api_client_stig_checklists.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: tests/test_homebrew_formula.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: tests/test_mcp_plans_handler.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: tests/test_mcp_systems_coverage.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: tests/test_plans.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: tests/test_preflight.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: tests/test_provisioning_engine.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: tests/test_resolvers.py: path traversal ('../') combined with file access discloses broad/destructive filesystem access: tests/test_scope_artifact_compose_recipe.py: destructive filesystem call (shutil.rmtree/os.remove/etc.) discloses broad/destructive filesystem access: tests/recipes/test_openscap_run_scan.py: path traversal ('../') combined with file access 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 0.26.11 (patch bump): new capability primitive(s) in source: child_process, dynamic-eval, fs-write, http-client, subprocess; new hardcoded outbound host/BCC destination(s): 127.0.0.1, alias.custom.example, api, api.example.com, api.openai.com, api.test, app.example.com, chatgpt.com, checklists.nist.gov, codex.legacy, configured.example, custom-openai.example, custom.example.com, default.example.com, developer.mozilla.org, different.com, docs.aws.amazon.com, docs.example.com, docs.pretorin.com, docs.sigstore.dev, env.example, env.test, example.com, example.test, from-config.example, garage, github.com, harness.legacy, in-toto.io, learn.microsoft.com, legacy-codex.example, legacy-harness.example, legacy-stored.example, legacy.example, llm.example, mcp.example.com, model-primary.example, model.configured, model.env, model.env.example, model.setter, model.test, models.example, models.example.com, my-vllm.example.com, omnitruck.chef.io, other.com, plat.test, platform.custom.example, platform.example, platform.pretorin.com, play.rust-lang.org, pretorin-models.example, pretorin.com, proxy.example, pypi.org, schemas.openxmlformats.org, self-hosted.example.com, slsa.dev, storage.example, stored-openai.example, test.pretorin.api, unknown.example, via-alias.example, wrong.example, www.w3.org.

• Transport config

  fail10% of grade

  Remote servers: TLS and auth mode (none/token/OAuth). Local servers: whether the manifest indicates it phones home.

  source performs runtime self-update / download-and-execute (bypasses version_behavior); manifest/source references tracking behavior — disclosure