urigen

AI Cost Tracking

🤖 LLM usage: $0.4211 (4 commits)
👤 Human dev: ~$272 (2.7h @ $100/h, 30min dedup)

Generated on 2026-06-15 using openrouter/qwen/qwen3-coder-next

urigen is the URI Ecosystem Generator. It creates proposal and ecosystem artifacts, then delegates validation to existing packages:

urigen generates.
touri registers capabilities.
uri2flow compiles flows.
uri3 explains and diagnoses.
uri2verify checks result contracts.
hypervisor deploys and manages lifecycle.

It is not a runtime transport and it does not execute backends directly.

CLI

urigen plan -p "stworz agenta pogodowego z healthcheckiem" --out output/proposals/weather.ecosystem.proposal.yaml
urigen generate output/proposals/weather.ecosystem.proposal.yaml --out output/ecosystems/weather
urigen verify output/ecosystems/weather/ecosystem.yaml
urigen explain output/ecosystems/weather/ecosystem.yaml

plan and verify are side-effect safe. generate writes only the requested output directory. apply --plan writes apply_plan.yaml and a diff preview. apply --approve mutates the repo transactionally, writes apply_result.json, and auto-rolls back on failure. Use --rollback to restore from rollback/manifest.json.

Current MVP

The first implementation is intentionally deterministic and repo-local:

plan      builds a proposal from a prompt and profile
generate  assembles a weather/voice ecosystem from existing repo artifacts
verify    validates capabilities, flows, sample URI explain output and doctor
explain   renders domains, agents, capabilities, flows, deployments and risks
apply     --plan shows diff; --approve executes ApplyPlan; failure auto-rolls back
          --rollback restores from rollback/manifest.json
schema    check validates ecosystem YAML envelopes and apply artifacts

urigen imports high-level validators from touri, uri2flow and uri3. It must not import uri2run.transports, uri2ops.server or low-level hypervisor process runtime modules.

Artifact Contract

Every generated ecosystem directory contains:

ecosystem.yaml
README.md
capabilities/
flows/
contracts/
deployments/
tests/

ecosystem.yaml is the source of truth for verification and explain output. Generated proposal and ecosystem YAML files are canonical URI3 artifacts: $schema, apiVersion, kind, metadata, uri.self and spec are emitted at the top of the document, while legacy-compatible fields remain available for existing readers.

Generated capability manifests receive a default, non-failing data_quality policy so uri3 explain and urigen explain can distinguish an intentionally checked generated path from an ungoverned capability.

Profiles

The planner accepts these profile names:

minimal — weather demo ecosystem
voice — weather + voice capabilities
dashboard-agent — hypervisor dashboard system agent (view/repair/ticket UI)
agent, operator, provider, full — extended profiles (partial)

Friendly aliases are accepted for onboarding:

dashboard → dashboard-agent
voice-agent → voice
operator-agent → operator
ecosystem / full-ecosystem → full

minimal
agent
voice
operator
provider
full
dashboard-agent

List profiles from the CLI:

urigen profiles
uri ecosystem profiles

The MVP uses minimal and voice paths directly. The other profiles are reserved in the proposal contract and should be expanded without moving runtime execution into urigen.

Verification

urigen verify checks:

touri manifest validation
uri2flow validation and workflow graph validation
uri3 explain for each sample_uri
uri3 doctor against the generated capability registry

When report writing is enabled, it writes verify_report.json next to ecosystem.yaml.

Next Work

JSON Schema validation on plan/generate
dashboard app/ generation inside ecosystem output
profile-specific generators (agent, operator, provider, full)
markpact README export with capability/flow/deploy/test blocks
recovery reports and uri2verify replay integration

Install

pip install -e .

License

Licensed under Apache-2.0.