@darkelogix/openclaw-trusted-mode
v1.0.7
Published
MIT-licensed OpenClaw Trusted Mode plugin with standalone hardening and optional SDE-backed governance
Maintainers
Readme
@darkelogix/openclaw-trusted-mode
Terminology and acronyms: GLOSSARY.md.
Acronym Expansions
SDE: Strategic Decision EnginePDP: Policy Decision PointWSL: Windows Subsystem for LinuxCI: Continuous Integration
OpenClaw plugin that enforces Trusted Mode policy checks on before_tool_call.
Documentation index (by audience and task): docs/README.md.
npm Package
Install the public MIT adapter/plugin package with:
npm install @darkelogix/openclaw-trusted-modeWhat npm install gives you
npm install @darkelogix/openclaw-trusted-mode gives you the MIT adapter/plugin layer and standalone hardening flow only. It does not grant access to the proprietary SDE runtime, enterprise deployment packs, or governed tenant entitlements.
Need governed mode?
If you want SDE-backed governed mode, obtain your licensed SDE runtime and deployment instructions from the Darkelogix customer console. Use the public npm package for adapter installation, then connect it to your licensed SDE environment for governed authorization, evidence, and rollout controls.
The npm package contains the MIT plugin files and standalone hardening logic only.
It does not include the proprietary sde-enterprise runtime.
Licensing
openclaw-trusted-mode is licensed under the MIT License.
sde-enterprise, including the SDE PDP runtime and related enterprise deployment assets, is proprietary software and is not covered by the plugin's MIT license. Use, copying, modification, distribution, or deployment of the SDE runtime requires a separate commercial license or written permission from Darkelogix.
First-time setup (download/install/configure/test/run): START_HERE.md.
Troubleshooting decision tree: SELF_SERVICE_FAQ.md.
Org defaults and support metadata: <org-values-file>.
One-command setup: powershell -ExecutionPolicy Bypass -File <bootstrap-self-service-script-path>.
For full install/reinstall/uninstall/startup/config/troubleshooting guidance across both plugin and SDE-PDP, see OPERATIONS_GUIDE.md.
For a simpler operator runbook, see RUNBOOK_NON_TECHNICAL.md.
For go-live gating, use PRODUCTION_READINESS_CHECKLIST.md.
For a pre-filled starting point, use PRODUCTION_READINESS_CHECKLIST_EXAMPLE.md.
For alternate port deployments, use PRODUCTION_READINESS_CHECKLIST_EXAMPLE_ALT_PORTS.md.
For public launch readiness, use PUBLIC_RELEASE_READINESS_CHECKLIST.md.
For end-to-end public release execution steps (what/where/how), use PUBLIC_RELEASE_PROCESS_RUNBOOK.md.
For certified runtime support status, see COMPATIBILITY_MATRIX.md.
For vulnerability reporting and security posture, see SECURITY.md.
For release hardening process, see RELEASE_OPERATIONS.md.
For security evidence indexing, see SECURITY_EVIDENCE_BUNDLE.md.
For performance baseline evidence, see PERFORMANCE_BASELINE.md.
For governed release declaration, see RELEASE_v1.0.0.md.
What it does
- Free standalone mode defaults to local hardening with a minimal allowlist:
read_filelist_filessearch_files
- Blocks high-risk tools such as
exec, file writes/edits, and deletes unless you deliberately widen the policy. - Sends tool call context to a Policy Decision Point (PDP) endpoint.
- Denies execution when PDP returns a deny decision.
- Optionally enforces returned constraints.
- Supports fail-closed (default) or fail-open behavior.
Free vs Paid
The product boundary should be explicit at install time:
npm installgets you the adapter/plugin and standalone hardening pathgoverned mode requires a separately licensed SDE deployment
the customer console is the supported way to obtain governed runtime artifacts and deployment instructions
Free standalone use:
- useful as a local hardening layer
- works without
sde-enterprise - best for "read/search only" OpenClaw sessions
Paid / enterprise use:
- PDP-backed authorization and deny decisions
- signed policy packs
- tenant entitlements and governed rollout
- release attestation and compatibility certification
Build and test
npm run build
npm test
npm run adversarial-check
npm run performance-benchmark
npm run test-pack-matrixTrusted Mode Check
npm run trusted-mode-check
npm run trusted-mode-check -- --jsontrusted-mode-check is a PDP-backed validation path. It is useful for SDE-integrated deployments, not for standalone free-mode validation.
JSON output status values:
ENFORCED_OKLOCKDOWN_ONLYUNSAFE
Attestation pack inputs:
attestation/trusted_mode_attest_v1.jsonattestation/trusted_mode_attest_v1.sig
Runtime/certification env vars:
CERTIFICATION_STATUS(CERTIFIED_ENFORCED|LOCKDOWN_ONLY|UNSUPPORTED)OPENCLAW_VERSIONEXPECTED_STATUS(optional CI assertion override)
Local install in OpenClaw (WSL)
openclaw plugins install /mnt/c/path/to/openclaw-trusted-mode
openclaw plugins info openclaw-trusted-modeFor a standalone free-mode config, start from openclaw.user-config.entry.example.json.
For governed mode, install/register the plugin first, then write the OpenClaw host config with:
openclaw-trusted-mode-configure \
--tenantId darkelogix \
--gatewayId gw-dev \
--environment dev \
--pdpUrl http://10.90.0.6:8001/v1/authorize \
--certificationStatus LOCKDOWN_ONLYThis command updates ~/.openclaw/openclaw.json, adds openclaw-trusted-mode to plugins.allow, and writes the governed plugin settings under plugins.entries.openclaw-trusted-mode.
Plugin config
See openclaw.plugin.json for config schema and defaults, including:
pdpUrlpolicyVariantpdpTimeoutMsfailClosedtenantIdcertificationStatusopenclawVersioncertifiedOpenClawVersionshighRiskToolstoolPolicyModeallowedToolsrequireTenantIdallowedTenantIdscontextCurator
Recommended standalone free-mode baseline:
{
"toolPolicyMode": "ALLOWLIST_ONLY",
"allowedTools": ["read_file", "list_files", "search_files"],
"failClosed": true,
"certificationStatus": "LOCKDOWN_ONLY"
}Recommended paid / PDP-backed baseline:
{
"toolPolicyMode": "PDP",
"pdpUrl": "http://localhost:8001/v1/authorize",
"tenantId": "trial-tenant",
"gatewayId": "gw-smoke-1",
"environment": "prod",
"failClosed": true,
"certificationStatus": "LOCKDOWN_ONLY"
}Compatibility Matrix Automation
npm run update-compatibility-matrix
npm run verify-compatibility-matrixSecurity Gates
npm run collect-security-evidence
npm run generate-security-release-index
npm run verify-security-gatesSchema Contract and Evidence Bundle
npm run verify-plugin-schema-contract
npm run bundle-release-evidenceStartup Health Verification
npm run startup-health-check -- --skip-plugin-check