Files
cp1-agents/CLAUDE.md
Jonathan Boniface 416c66ced4
All checks were successful
check agents / check (push) Successful in 10s
Add CP1 Discovery agents in both patterns
ALZ-F01 Discovery as a sub-agent orchestrator (sub-agents/) and as eight
individual declarative agents (individual-agents/), with their KAB workflow
exports. Shared check, render and import scripts, the alz-discovery-sample
MCP server, the crucible-kab-dev deploy profile, deployment docs, platform
issue reports and the source material. Agent steps time out at 20 minutes.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 18:25:57 +00:00

2.6 KiB

cp1-agents

KAB agents and workflows for the CP1 Agentic Landing Zone Epic, in two patterns. Modelled on cp4-agents.

@.agents/CONTEXT.md

Each Feature folder has its own CLAUDE.md, and so does each agent folder. Read them before changing anything inside.

Commands (repository root)

find sub-agents individual-agents -mindepth 3 -maxdepth 3 -name agent.yaml -print0 | xargs -0 -n1 dirname -z \
  | xargs -0 python3 scripts/check_agents.py --runtime-image cruciblecoredevacr.azurecr.io/kab/agent-runtime-adk:0.35.7
python3 "<pattern>/<Feature>/scripts/build_workflow.py"                     # after any schemas/ change
python3 scripts/render_k8s.py deploy/crucible-kab-dev.yaml [<component>]   # fill placeholders into build/
kustomize build build/crucible-kab-dev/<component>                          # validate a render
python3 scripts/kab_import_workflow.py "<pattern>/<Feature>" --project-id <id>   # validate; --apply to import

Rules

  • Keep the two patterns apart. sub-agents/ and individual-agents/ never share agent code, prompts or ADRs. They share only scripts/, mcp-servers/, deploy/ and docs/.
  • Don't duplicate. A script, server or doc that more than one Feature or pattern needs belongs at the root, once. Per-agent Dockerfile and kubernetes/ are the exception, because Kyndryl CI needs them in every agent folder.
  • Keep the kubernetes/ placeholders (<placeholder_for_*>). Fill them only in build/ (git-ignored), through render_k8s.py.
  • No key, token or password in any file, including deploy/ profiles, docs and examples. Keys live in Kubernetes Secrets created by hand.
  • .env files are local only. They're git-ignored (.gitignore); never commit one, never force-add one, and never commit an archive that contains one. docs/source-material/ is read-only input.
  • Cluster changes (kubectl apply/delete/create, image pushes) need the user's explicit approval each time. kubectl diff, get and kustomize build don't. Never read Secret values.
  • Never kubectl delete -k an agent's own kubernetes/ folder. Its namespace.yaml is the shared namespace.
  • Don't change a deployed agent's repo label in deploy/: it's in the Deployment's selector.
  • Human docs go in README.md and docs/; agent context goes in CLAUDE.md and .agents/.
  • A change to an original requirement (the ADO story or the v1.1 story file) needs an ADR in that pattern's ADRs and a step in its .agents/PLAN.md.
  • Never change existing ADO work items from this repository's work.
  • Quote the Feature paths. They contain spaces and :. In scripts, pass them NUL-separated (-print0, xargs -0).