Keys and possible values
Keys and possible values
Section titled “Keys and possible values”This page is generated from the arxo metrics registry --registry-format detailed schema for OpenClaw Architecture (openclaw_architecture).
Detector keys
Section titled “Detector keys”| Result key | Axis | Rule ID | Impact | Effort | Score range | Severity values | Recommendation |
|---|---|---|---|---|---|---|---|
openclaw_architecture.soul_evil_hook_enabled_score | ConfigSecurity | arxo/openclaw-soul-evil-hook-enabled | critical | low | 0..1 | Good, Low, Medium, High, Critical | Disable soul-evil hook and remove SOUL_EVIL.md; prevent silent persona swap. |
openclaw_architecture.config_patch_unrestricted_score | ConfigSecurity | arxo/openclaw-config-patch-unrestricted | critical | low | 0..1 | Good, Low, Medium, High, Critical | Add config.patch to restricted_actions or tools.deny to prevent agent self-modification. |
openclaw_architecture.hook_whitelist_absence_score | ConfigSecurity | arxo/openclaw-hook-whitelist-absence | high | medium | 0..1 | Good, Low, Medium, High, Critical | Configure hooks.allowed allowlist when using custom hooks. |
openclaw_architecture.webhook_auth_absence_score | ConfigSecurity | arxo/openclaw-webhook-auth-absence | high | medium | 0..1 | Good, Low, Medium, High, Critical | Enable webhook signature verification for Twilio/Telnyx; set skipSignatureVerification false. |
openclaw_architecture.browser_domain_allowlist_absence_score | ConfigSecurity | arxo/openclaw-browser-domain-allowlist-absence | high | medium | 0..1 | Good, Low, Medium, High, Critical | Set browser.domainAllowlist when browser tool is enabled to limit navigation. |
openclaw_architecture.contact_allowlist_absence_score | ConfigSecurity | arxo/openclaw-contact-allowlist-absence | high | low | 0..1 | Good, Low, Medium, High, Critical | Configure contactAllowlist for channels to restrict who can message the agent. |
openclaw_architecture.sensitive_directory_exposure_score | ConfigSecurity | arxo/openclaw-sensitive-directory-exposure | high | medium | 0..1 | Good, Low, Medium, High, Critical | Block ~/.ssh, ~/.aws, ~/.config, ~/.gnupg, ~/.kube, /etc in sandbox paths. |
openclaw_architecture.credential_context_window_exposure_score | ConfigSecurity | arxo/openclaw-credential-context-window-exposure | high | medium | 0..1 | Good, Low, Medium, High, Critical | Use env vars or vault for API keys when shell/browser tools are enabled. |
openclaw_architecture.unbounded_tool_access_score | SkillGovernance | arxo/openclaw-unbounded-tool-access | critical | medium | 0..1 | Good, Low, Medium, High, Critical | Restrict agents.list[].tools.allow to explicit list; avoid wildcard. |
openclaw_architecture.max_steps_absence_score | SkillGovernance | arxo/openclaw-max-steps-absence | high | low | 0..1 | Good, Low, Medium, High, Critical | Set maxSteps or maxIterations for agents to bound execution. |
openclaw_architecture.cascade_depth_unlimited_score | SkillGovernance | arxo/openclaw-cascade-depth-unlimited | high | medium | 0..1 | Good, Low, Medium, High, Critical | Set agents.maxCascadeDepth when using multiple agents. |
openclaw_architecture.read_only_triage_absence_score | SkillGovernance | arxo/openclaw-read-only-triage-absence | medium | high | 0..1 | Good, Low, Medium, High, Critical | Use read-only triage agent in front of tool-enabled agent for untrusted input. |
openclaw_architecture.inter_agent_auth_absence_score | SkillGovernance | arxo/openclaw-inter-agent-auth-absence | high | medium | 0..1 | Good, Low, Medium, High, Critical | Configure inter-agent auth when using multiple agents in agents.list. |
openclaw_architecture.agent_output_sanitization_gap_score | SkillGovernance | arxo/openclaw-agent-output-sanitization-gap | high | medium | 0..1 | Good, Low, Medium, High, Critical | Enable output.sanitize or equivalent when agent uses shell/DB/API tools. |
openclaw_architecture.sub_agent_privilege_inheritance_score | SkillGovernance | arxo/openclaw-sub-agent-privilege-inheritance | high | medium | 0..1 | Good, Low, Medium, High, Critical | Set per-agent tools.allow for sub-agents; avoid full inheritance. |
openclaw_architecture.prompt_injection_defense_absence_score | SkillGovernance | arxo/openclaw-prompt-injection-defense-absence | high | medium | 0..1 | Good, Low, Medium, High, Critical | Configure security.promptInjection or equivalent for input scanning and anchoring. |
openclaw_architecture.external_content_scan_absence_score | SkillGovernance | arxo/openclaw-external-content-scan-absence | high | medium | 0..1 | Good, Low, Medium, High, Critical | Enable tools.contentScan or equivalent when using browser/email/file tools. |
openclaw_architecture.self_replication_capability_score | SkillGovernance | arxo/openclaw-self-replication-capability | critical | low | 0..1 | Good, Low, Medium, High, Critical | Set agents.selfReplication to deny when agent has shell and network access. |
openclaw_architecture.shutdown_mechanism_absence_score | SkillGovernance | arxo/openclaw-shutdown-mechanism-absence | high | low | 0..1 | Good, Low, Medium, High, Critical | Enable interruptible or killSwitch for agents to allow human halt. |
openclaw_architecture.goal_drift_detection_absence_score | Observability | arxo/openclaw-goal-drift-detection-absence | medium | medium | 0..1 | Good, Low, Medium, High, Critical | Configure diagnostics.drift_detection or goal_monitoring when agents have long sessions. |
openclaw_architecture.memory_file_encryption_absence_score | SupplyChain | arxo/openclaw-memory-file-encryption-absence | high | medium | 0..1 | Good, Low, Medium, High, Critical | Enable memory.encryption when memory is writable to protect IDENTITY/SOUL on disk. |
openclaw_architecture.identity_file_mutable_by_agent_score | SupplyChain | arxo/openclaw-identity-file-mutable-by-agent | high | medium | 0..1 | Good, Low, Medium, High, Critical | Add integrity/checksum for IDENTITY.md and SOUL.md when agent can write memory. |
openclaw_architecture.hardcoded_credential_in_config_score | SupplyChain | arxo/openclaw-hardcoded-credential-in-config | critical | medium | 0..1 | Good, Low, Medium, High, Critical | Use env vars or vault for API keys; remove hardcoded credentials from config. |
openclaw_architecture.network_egress_unrestricted_score | SupplyChain | arxo/openclaw-network-egress-unrestricted | high | medium | 0..1 | Good, Low, Medium, High, Critical | Set network.egress.allowlist when sandbox has bridge or host network access. |
openclaw_architecture.a2a_session_smuggling_risk_score | SkillGovernance | arxo/openclaw-a2a-session-smuggling-risk | critical | medium | 0..1 | Good, Low, Medium, High, Critical | Enable AgentCard verification or context grounding for A2A to prevent session smuggling. |
openclaw_architecture.a2a_unverified_agent_card_with_tool_access_score | SupplyChain | arxo/openclaw-a2a-unverified-agent-card-with-tool-access | critical | medium | 0..1 | Good, Low, Medium, High, Critical | Enable AgentCard verification for A2A or restrict tool access when using unverified remotes. |
openclaw_architecture.agent_identity_registry_absence_score | ConfigSecurity | arxo/openclaw-agent-identity-registry-absence | high | medium | 0..1 | Good, Low, Medium, High, Critical | Configure agents.registry or per-agent owner/purpose/id for attribution. |
openclaw_architecture.agent_tool_over_permission_score | SkillGovernance | arxo/openclaw-agent-tool-over-permission | high | medium | 0..1 | Good, Low, Medium, High, Critical | Restrict agent tool permissions. |
openclaw_architecture.audit_log_disabled_score | Observability | arxo/openclaw-audit-log-disabled | high | low | 0..1 | Good, Low, Medium, High, Critical | Enable audit log for diagnostics. |
openclaw_architecture.behavioral_drift_detection_absence_score | Observability | arxo/openclaw-behavioral-drift-detection-absence | medium | medium | 0..1 | Good, Low, Medium, High, Critical | Enable behavioral drift detection. |
openclaw_architecture.clawhub_skill_provenance_score | SupplyChain | arxo/openclaw-clawhub-skill-provenance | high | medium | 0..1 | Good, Low, Medium, High, Critical | Use ClawHub or verified skill sources. |
openclaw_architecture.content_capture_disabled_score | Observability | arxo/openclaw-content-capture-disabled | medium | low | 0..1 | Good, Low, Medium, High, Critical | Set diagnostics.otel.captureContent: true. |
openclaw_architecture.control_ui_exposure_score | ConfigSecurity | arxo/openclaw-control-ui-exposure | high | medium | 0..1 | Good, Low, Medium, High, Critical | Restrict control UI when gateway is remote. |
openclaw_architecture.cost_tracking_absence_score | Observability | arxo/openclaw-cost-tracking-absence | medium | low | 0..1 | Good, Low, Medium, High, Critical | Enable cost tracking for providers. |
openclaw_architecture.covert_tool_invocation_instruction_score | SupplyChain | arxo/openclaw-covert-tool-invocation-instruction | high | medium | 0..1 | Good, Low, Medium, High, Critical | Remove covert tool invocation instructions from SKILL.md. |
openclaw_architecture.credential_harvesting_phrase_score | SupplyChain | arxo/openclaw-credential-harvesting-phrase | critical | low | 0..1 | Good, Low, Medium, High, Critical | Remove credential harvesting phrases from SKILL.md. |
openclaw_architecture.dangerous_skill_exposure_score | ConfigSecurity | arxo/openclaw-dangerous-skill-exposure | high | medium | 0..1 | Good, Low, Medium, High, Critical | Restrict dangerous skills or require approval. |
openclaw_architecture.discovery_pinning_gap_score | ConfigSecurity | arxo/openclaw-discovery-pinning-gap | critical | medium | 0..1 | Good, Low, Medium, High, Critical | Pin gateway address when using mDNS/DNS-SD discovery. |
openclaw_architecture.dm_policy_open_score | ConfigSecurity | arxo/openclaw-dm-policy-open | high | low | 0..1 | Good, Low, Medium, High, Critical | Use dmPolicy pairing or allowlist; avoid open for unknown senders. |
openclaw_architecture.egress_firewall_absence_score | SupplyChain | arxo/openclaw-egress-firewall-absence | high | medium | 0..1 | Good, Low, Medium, High, Critical | Configure egress allowlist when sandbox is on. |
openclaw_architecture.exec_security_full_mode_score | SkillGovernance | arxo/openclaw-exec-security-full-mode | high | medium | 0..1 | Good, Low, Medium, High, Critical | Avoid exec security full mode or allowlist only. |
openclaw_architecture.external_binary_install_instruction_score | SupplyChain | arxo/openclaw-external-binary-install-instruction | high | medium | 0..1 | Good, Low, Medium, High, Critical | Avoid instructing users to download and run binaries. |
openclaw_architecture.fake_prerequisite_in_skill_md_score | SupplyChain | arxo/openclaw-fake-prerequisite-in-skill-md | high | medium | 0..1 | Good, Low, Medium, High, Critical | Remove fake prerequisites from SKILL.md. |
openclaw_architecture.gateway_auth_gap_score | ConfigSecurity | arxo/openclaw-gateway-auth-gap | critical | low | 0..1 | Good, Low, Medium, High, Critical | Set gateway.auth.token when gateway.remote is true. |
openclaw_architecture.gen_ai_semantic_convention_gap_score | Observability | arxo/openclaw-gen-ai-semantic-convention-gap | medium | low | 0..1 | Good, Low, Medium, High, Critical | Use GenAI semantic conventions for telemetry. |
openclaw_architecture.high_risk_source_without_isolation_score | SupplyChain | arxo/openclaw-high-risk-source-without-isolation | high | medium | 0..1 | Good, Low, Medium, High, Critical | Use sandbox and egress allowlist for high-risk sources. |
openclaw_architecture.human_approval_gap_score | SkillGovernance | arxo/openclaw-human-approval-gap | high | medium | 0..1 | Good, Low, Medium, High, Critical | Configure approval workflow for dangerous skills. |
openclaw_architecture.installed_skill_fails_fake_prerequisite_check_score | SupplyChain | arxo/openclaw-installed-skill-fails-fake-prerequisite-check | high | medium | 0..1 | Good, Low, Medium, High, Critical | Ensure installed skills pass prerequisite checks. |
openclaw_architecture.interruptibility_gap_score | Observability | arxo/openclaw-interruptibility-gap | medium | low | 0..1 | Good, Low, Medium, High, Critical | Set agents.interruptible where appropriate. |
openclaw_architecture.known_malicious_skill_id_score | SupplyChain | arxo/openclaw-known-malicious-skill-id | critical | low | 0..1 | Good, Low, Medium, High, Critical | Remove known malicious skill IDs from config. |
openclaw_architecture.log_retention_too_short_score | Observability | arxo/openclaw-log-retention-too-short | medium | low | 0..1 | Good, Low, Medium, High, Critical | Set log retention to 90 days or more. |
openclaw_architecture.long_lived_credential_per_agent_score | ConfigSecurity | arxo/openclaw-long-lived-credential-per-agent | high | medium | 0..1 | Good, Low, Medium, High, Critical | Use short-lived scoped tokens or set credentials.ttl/refresh for agents. |
openclaw_architecture.loopback_binding_gap_score | ConfigSecurity | arxo/openclaw-loopback-binding-gap | high | medium | 0..1 | Good, Low, Medium, High, Critical | Bind to loopback when not using remote gateway. |
openclaw_architecture.mcp_descriptor_mutation_risk_score | SkillGovernance | arxo/openclaw-mcp-descriptor-mutation-risk | high | medium | 0..1 | Good, Low, Medium, High, Critical | Pin MCP tool descriptors (descriptionHash, schemaHash, or pinDescriptors) to prevent rug pulls. |
openclaw_architecture.memory_file_write_risk_score | SupplyChain | arxo/openclaw-memory-file-write-risk | high | medium | 0..1 | Good, Low, Medium, High, Critical | Restrict file write access for skills. |
openclaw_architecture.memory_injection_phrase_score | SupplyChain | arxo/openclaw-memory-injection-phrase | high | medium | 0..1 | Good, Low, Medium, High, Critical | Remove memory-injection style instructions from SKILL.md. |
openclaw_architecture.metadata_validation_gap_score | SupplyChain | arxo/openclaw-metadata-validation-gap | medium | medium | 0..1 | Good, Low, Medium, High, Critical | Validate skill metadata and checksums. |
openclaw_architecture.nonexistent_official_dependency_score | SupplyChain | arxo/openclaw-nonexistent-official-dependency | high | low | 0..1 | Good, Low, Medium, High, Critical | Use official package names in prerequisites. |
openclaw_architecture.otel_plugin_absence_score | Observability | arxo/openclaw-otel-plugin-absence | high | low | 0..1 | Good, Low, Medium, High, Critical | Enable diagnostics-otel plugin. |
openclaw_architecture.pastebin_piping_in_skill_md_score | SupplyChain | arxo/openclaw-pastebin-piping-in-skill-md | high | medium | 0..1 | Good, Low, Medium, High, Critical | Avoid pastebin/rentry piping in SKILL.md. |
openclaw_architecture.plaintext_credential_storage_score | ConfigSecurity | arxo/openclaw-plaintext-credential-storage | critical | medium | 0..1 | Good, Low, Medium, High, Critical | Avoid storing credentials in plaintext. |
openclaw_architecture.prerequisite_section_present_without_allowlist_score | SupplyChain | arxo/openclaw-prerequisite-section-present-without-allowlist | medium | low | 0..1 | Good, Low, Medium, High, Critical | Configure skills.prerequisiteAllowlist when using prerequisites. |
openclaw_architecture.prompt_injection_in_skill_md_score | SupplyChain | arxo/openclaw-prompt-injection-in-skill-md | high | medium | 0..1 | Good, Low, Medium, High, Critical | Avoid prompt injection patterns in SKILL.md. |
openclaw_architecture.rag_knowledge_provenance_gap_score | SupplyChain | arxo/openclaw-rag-knowledge-provenance-gap | high | medium | 0..1 | Good, Low, Medium, High, Critical | Enable provenance tracking or source validation for knowledge/retrieval. |
openclaw_architecture.rate_limit_absence_score | ConfigSecurity | arxo/openclaw-rate-limit-absence | medium | medium | 0..1 | Good, Low, Medium, High, Critical | Configure rate limiting for tool/API calls. |
openclaw_architecture.reasoning_trace_capture_absence_score | Observability | arxo/openclaw-reasoning-trace-capture-absence | medium | medium | 0..1 | Good, Low, Medium, High, Critical | Configure diagnostics.reasoningTrace or chainOfThoughtCapture for decision visibility. |
openclaw_architecture.reverse_shell_or_c2_indicator_score | SupplyChain | arxo/openclaw-reverse-shell-or-c2-indicator | critical | low | 0..1 | Good, Low, Medium, High, Critical | Remove reverse shell or C2 indicators from SKILL.md. |
openclaw_architecture.sample_rate_misconfigured_score | Observability | arxo/openclaw-sample-rate-misconfigured | medium | low | 0..1 | Good, Low, Medium, High, Critical | Configure sample rate for diagnostics. |
openclaw_architecture.sandbox_absence_score | SkillGovernance | arxo/openclaw-sandbox-absence | high | medium | 0..1 | Good, Low, Medium, High, Critical | Enable sandbox for dangerous tools. |
openclaw_architecture.self_modifying_skill_risk_score | SkillGovernance | arxo/openclaw-self-modifying-skill-risk | high | medium | 0..1 | Good, Low, Medium, High, Critical | Gate dynamic Skill.create with user content. |
openclaw_architecture.shadow_mcp_server_score | SkillGovernance | arxo/openclaw-shadow-mcp-server | medium | medium | 0..1 | Good, Low, Medium, High, Critical | Use known MCP server sources only. |
openclaw_architecture.signed_system_prompt_absence_score | ConfigSecurity | arxo/openclaw-signed-system-prompt-absence | high | medium | 0..1 | Good, Low, Medium, High, Critical | Add system prompt signing or integrity to prevent goal hijacking. |
openclaw_architecture.skill_allowlist_absence_score | SkillGovernance | arxo/openclaw-skill-allowlist-absence | critical | medium | 0..1 | Good, Low, Medium, High, Critical | Configure skills.allowed and/or skills.denied. |
openclaw_architecture.skill_privilege_escalation_phrase_score | SupplyChain | arxo/openclaw-skill-privilege-escalation-phrase | high | medium | 0..1 | Good, Low, Medium, High, Critical | Remove privilege escalation instructions from SKILL.md. |
openclaw_architecture.skill_source_integrity_score | SupplyChain | arxo/openclaw-skill-source-integrity | high | medium | 0..1 | Good, Low, Medium, High, Critical | Use checksums or signed skill sources. |
openclaw_architecture.skill_virus_scan_absence_score | SupplyChain | arxo/openclaw-skill-virus-scan-absence | high | medium | 0..1 | Good, Low, Medium, High, Critical | Enable virus scan for skill sources. |
openclaw_architecture.tool_poisoning_susceptibility_score | SkillGovernance | arxo/openclaw-tool-poisoning-susceptibility | high | medium | 0..1 | Good, Low, Medium, High, Critical | Restrict MCP tools and sandbox when needed. |
openclaw_architecture.typosquat_skill_name_score | SupplyChain | arxo/openclaw-typosquat-skill-name | high | medium | 0..1 | Good, Low, Medium, High, Critical | Use official skill names; avoid typosquatting. |
openclaw_architecture.unpinned_mcp_server_version_score | SupplyChain | arxo/openclaw-unpinned-mcp-server-version | high | low | 0..1 | Good, Low, Medium, High, Critical | Pin MCP server version; avoid latest or *. |
openclaw_architecture.weak_auth_token_score | ConfigSecurity | arxo/openclaw-weak-auth-token | high | low | 0..1 | Good, Low, Medium, High, Critical | Use 256-bit random token for gateway auth. |
openclaw_architecture.windows_password_protected_zip_in_prereq_score | SupplyChain | arxo/openclaw-windows-password-protected-zip-in-prereq | high | medium | 0..1 | Good, Low, Medium, High, Critical | Avoid password-protected zip in prerequisites. |
Possible values
Section titled “Possible values”- Score:
0.0= issue detected,1.0= healthy. Per-detector scores are combined into axis and overall health. - Severity:
Good,Low,Medium,High,Critical(from detectorMetricScore).