Docs
Contributing
Local development
Concise path from clone to a working audit service, documentation gates, and tests. For product semantics, stay with the canonical guides linked at the end.
Prerequisites#
- Git, Docker with Compose v2, Python 3.10+, Rust toolchain (
cargo). - Ports 8088 (audit HTTP) and 5432 (Postgres) available locally.
1. Clone#
2. Python tooling (venv + editable install)#
The CLI and tests live under python/. Create a venv inside python/ (matches Makefile targets that run cd python && . .venv/bin/activate):
If pytest fails at import time with ModuleNotFoundError (for example nacl), install the editable package in that venv as above; PyNaCl is declared in python/pyproject.toml.
3. Audit service via Docker Compose#
From the repository root:
Health (liveness; available only after the service has bound HTTP):
curl -fsS http://127.0.0.1:8088/health
Operational readiness (Postgres + migrations + writable ledger expectations used by the service):
Root docker-compose.yml sets GOVAI_API_KEYS for the container (see file for the dev value). Use the same bearer token when calling authenticated routes from your shell.
4. Optional root .env#
Makefile does -include .env. Copy .env.example to .env for local overrides only; never commit secrets.
5. Documentation report gate#
From the repository root (uses system python3 and scripts/gate_reports.py):
5b. OSS health, metrics, and local link checks (stdlib scripts)#
These targets use stdlib-only Python under scripts/ (no network, no writes):
Machine-readable JSON (automation / CI mirrors):
CI (.github/workflows/oss-developer-experience.yml) uploads structured diagnostics as the oss-check-json artifact after make enterprise-readiness-check. Browse the catalog interactively:
CI artifact explorer
70 files uploaded as oss-check-json after make enterprise-readiness-check. Search or expand a domain to inspect outputs.
Repository health
Required OSS files, layout checks, and contributor onboarding paths.
repo-health.jsonOSS health gate outputjson
Trust and security
Security-trust diagnostics and machine-readable trust manifest validation.
security-trust.jsonStructured security/trust checksjsontrust-manifest.jsonPublished trust manifest snapshotjsontrust-manifest-validation.jsonManifest schema validationjsontrust-chain.jsonImmutable trust chain validationjson
Documentation quality
Strict link checks and aggregated OSS diagnostics.
docs-links.jsonBroken local link reportjsonoss-diagnostics.jsonAggregated OSS diagnostics JSONjson
Commercial readiness
Billing, pricing, deployment, and enterprise onboarding path checks.
commercial-readiness.jsonCommercial platform readinessjson
Pilot execution
Sales and pilot package validation for enterprise evaluations.
pilot-execution.jsonPilot execution diagnosticsjsonpilot-manifest-validation.jsonPilot manifest validationjson
Operations
Customer operations manifests and production readiness checklist.
customer-operations.jsonCustomer operations diagnosticsjsoncustomer-operations-manifest-validation.jsonOperations manifest validationjsonproduction-readiness-checklist.mdGenerated production readiness checklistmarkdown
Partner ecosystem
Partner program manifests and certification package smoke output.
partner-ecosystem.jsonPartner ecosystem diagnosticsjsonpartner-ecosystem-manifest-validation.jsonPartner manifest validationjsonpartner-certification-package.mdCertification package generator outputmarkdown
Regulatory evidence
EU AI Act obligations, regulatory manifests, and export reports.
regulatory-evidence.jsonRegulatory evidence diagnosticsjsonregulatory-manifest-validation.jsonRegulatory manifest validationjsonai-act-obligations-validation.jsonAI Act obligations validationjsonregulatory-evidence-export.mdRegulatory evidence Markdown exportmarkdown
Observability
Operational snapshots, health scores, and intelligence reports.
observability.jsonObservability diagnosticsjsonobservability-manifest-validation.jsonObservability manifest validationjsonoperational-snapshot-validation.jsonOperational snapshot validationjsonoperational-health-score.jsonOperational health scorejsonoperational-intelligence-report.mdOperational intelligence reportmarkdown
Runtime safety
Runtime guardrails, snapshots, scores, and safety reports.
runtime-safety.jsonRuntime safety diagnosticsjsonruntime-safety-manifest-validation.jsonRuntime safety manifest validationjsonruntime-safety-snapshot-validation.jsonRuntime safety snapshot validationjsonruntime-safety-score.jsonRuntime safety scorejsonruntime-safety-report.mdRuntime safety reportmarkdown
Evidence quality
Dataset provenance, evidence quality scoring, and governance reports.
evidence-quality.jsonEvidence quality diagnosticsjsonevidence-quality-manifest-validation.jsonEvidence quality manifest validationjsondataset-provenance-snapshot-validation.jsonDataset provenance snapshot validationjsonevidence-quality-score.jsonEvidence quality scorejsondataset-governance-report.mdDataset governance reportmarkdown
Policy intelligence
Governance control snapshots, coverage scores, and control reports.
policy-intelligence.jsonPolicy intelligence diagnosticsjsonpolicy-intelligence-manifest-validation.jsonPolicy intelligence manifest validationjsongovernance-control-snapshot-validation.jsonGovernance control snapshot validationjsonpolicy-coverage-score.jsonPolicy coverage scorejsongovernance-control-report.mdGovernance control reportmarkdown
Model risk
Model evaluation snapshots, risk scores, and assurance reports.
model-risk.jsonModel risk diagnosticsjsonmodel-risk-manifest-validation.jsonModel risk manifest validationjsonmodel-evaluation-snapshot-validation.jsonModel evaluation snapshot validationjsonmodel-risk-score.jsonModel risk scorejsonmodel-assurance-report.mdModel assurance reportmarkdown
Agent governance
Agent delegation snapshots, governance scores, and agent reports.
agent-governance.jsonAgent governance diagnosticsjsonagent-governance-manifest-validation.jsonAgent governance manifest validationjsonagent-delegation-snapshot-validation.jsonAgent delegation snapshot validationjsonagent-governance-score.jsonAgent governance scorejsonagent-governance-report.mdAgent governance reportmarkdown
Marketplace
Extension packages, policy packs, and marketplace listings.
marketplace.jsonMarketplace diagnosticsjsonmarketplace-manifest-validation.jsonMarketplace manifest validationjsonpolicy-pack-marketplace-manifest-validation.jsonPolicy pack marketplace manifestjsonpolicy-pack-eu-ai-act-basic-validation.jsonEU AI Act basic pack validationjsonpolicy-pack-internal-model-risk-validation.jsonInternal model risk pack validationjsonpolicy-pack-vendor-evaluation-validation.jsonVendor evaluation pack validationjsonextension-package-validation.jsonExtension package validationjsonmarketplace-listing.mdMarketplace listing outputmarkdown
Customer analytics
Customer health scoring and executive business review exports.
customer-analytics.jsonCustomer analytics diagnosticsjsoncustomer-analytics-manifest-validation.jsonCustomer analytics manifest validationjsoncustomer-health-score.jsonCustomer health scorejsonexecutive-business-review.mdExecutive business review reportmarkdown
Developer integrations
Automation packs, integration manifests, and summary reports.
developer-integrations.jsonDeveloper integrations diagnosticsjsondeveloper-integrations-manifest-validation.jsonIntegrations manifest validationjsonautomation-pack-validation.jsonAutomation pack validationjsonautomation-pack-summary.mdAutomation pack summarymarkdown
Release readiness
Changelog validation, release manifests, and readiness reports.
release-manifest-validation.jsonRelease manifest validationjsonchangelog-validation.jsonChangelog validationjsonrelease-readiness-report.jsonRelease readiness aggregate reportjsonrelease-notes-template.mdRelease notes templatemarkdown
Enterprise security review example (no audit service): bash examples/security-review/run-security-review-check.sh (from repo root).
Fail-closed demo (requires running audit stack + python/.venv with govai):
Read-only vs fail-closed semantics and exit codes: examples/local-demo/CONTRACT.md. Public docs preview: run the dashboard/ dev server (cd dashboard && npm ci && npm run dev) and open /docs / /help (content is read from ../docs/ at build/runtime).
Local audit read-only demo (optional)#
With the audit API already listening (for example after docker compose up -d --build):
These targets do not submit evidence, do not require API keys, and do not mutate the ledger. They are not part of make oss-diagnostics because a running service is not guaranteed in CI sandboxes.
Public documentation (dashboard)#
Canonical long-form documentation stays in docs/. The production-shaped reader UI is the dashboard/ App Router (app/docs, app/help). Canonical prose is not duplicated in a second Mintlify tree in this repository.
6. Rust tests#
7. Python tests#
With the python/.venv activated:
If you run python3 -m pytest python/tests from the repo root without the venv, collection may fail on missing optional deps — use the venv from step 2.
8. Optional: audit binary in background (Makefile)#
Canonical references#
- Quickstart narrative: quickstart-5min.md
- Golden path / evidence pack: golden-path.md, evidence-pack.md
- GitHub Action (artefact-bound CI): github-action.md
- Runtime evaluate (semantics, advisory vs enforcement): governance/runtime_integration.md, OpenAPI
api/govai-http-v1.openapi.yaml