Problem. No example demonstrates the project's actual value proposition end to end.
Proposed change. A workflow where an agent: reads a (fixture) alert, queries (mock) logs,
opens a (mock) issue, drafts a status update, and cannot restart the service without
approval. Tools are declared on the agent (exercising Epic A). The restart tool is in
policy.approvals.requiredFor. Ship a README narrative that walks validate → plan → apply → run (denied) → run --approve (allowed) → logs → audit verify.
Affected files. new examples/incident-triage/ (project.yaml, agents, tools, workflows,
policies, schemas, fixtures, README.md).
Acceptance criteria. Runs offline on the mock provider in CI; the un-approved run is blocked
with exit 5; the approved run completes; audit verify passes.
Test requirements. Add to the end-to-end example test sweep (mirror test/integration
patterns).
Depends on. #160, #161
Problem. No example demonstrates the project's actual value proposition end to end.
Proposed change. A workflow where an agent: reads a (fixture) alert, queries (mock) logs,
opens a (mock) issue, drafts a status update, and cannot restart the service without
approval. Tools are declared on the agent (exercising Epic A). The restart tool is in
policy.approvals.requiredFor. Ship a README narrative that walksvalidate → plan → apply → run(denied) →run --approve(allowed) →logs→audit verify.Affected files. new
examples/incident-triage/(project.yaml, agents, tools, workflows,policies, schemas, fixtures, README.md).
Acceptance criteria. Runs offline on the mock provider in CI; the un-approved run is blocked
with exit 5; the approved run completes; audit verify passes.
Test requirements. Add to the end-to-end example test sweep (mirror
test/integrationpatterns).
Depends on. #160, #161