- Critic 0.2.0 / 0.2.1: K-04 (risk), K-05 (confidence), K-06 (TCO), K-07 (routing) and K-11 (schema and 7R enum) run in code in the critic's hooks (app/arithmetic.py, app/business_logic.py); post_process sets those checks, the rejected dispositions, the sign-off queue and the verdict. The model keeps the judgement checks; a different reading of a code check is kept in unverified_claims. K-07 checks the disposition agent's own queue. Tests on the records of run exec_4d88027a4d26, whose model sums were wrong. - A represents edge is not a dependency (ADR 0012): risk scorer 0.1.1 and disposition 0.1.2 leave it out of dependents and dependencies; the disposition and critic prompts carry the risk scorer's rule. Run exec_1f518b04d6e4: APP-001's risk 1.00 -> 0.00, out of the tail. - The scaffold declares the critic's hooks, copies app/ in its Dockerfile, and keeps its app/ and tests/ when regenerating. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
CP1 - Assessment Critic & Sign-off Router Agent
Independently validates the Assessment: coverage, evidence, dependents, recalculated scores, confidence and TCO, the interim and advisory rules, and routes confidence below 0.80 to disposition sign-off.
ALZ-ASM-06 (ADO 3214464): one step of the workflow "CP1 - Agentic Landing Zone - Assessment". Version 0.1.0,
built 2026-11-02 on the KAB agent runtime 0.35.7 (agent-runtime-adk). The runtime builds the agent from agent.yaml:
its instructions are prompt.md, and it accepts and returns only what schemas/ declares. There is no
orchestrator; the workflow passes each step's output to the next.
Inputs
| Field | Type | Required | Description |
|---|---|---|---|
engagement_id |
string | yes | The engagement, e.g. ENG-MERIDIAN-001. |
run_id |
string | no | The Assessment run; the same id across the workflow's steps. |
discovery_execution_id |
string | yes | The Discovery run's execution id, e.g. exec_0d5c0e2a7f01. Every reference to a Discovery record names it. |
discovery_documents |
string / null | no | The Discovery run's step outputs as text: the Artifacts step's records, parsed by the parse step (p1.parsed_content). One JSON document per selected artifact, separated by a line '---'. |
discovery_files |
array / null | no | The parse step's file_metadata (p1.file_metadata): the selected artifacts' names, in the order their text appears in discovery_documents. |
rate_card_document |
string / null | no | The rate card's JSON, given as text in the run input rate_card_document (ADR 0009); see the Feature README. |
applications |
array | yes | |
tasks |
array | yes | |
coverage_gaps |
array | no | |
preconditions |
object | no | |
profiles |
array | no | |
profile_gaps |
array | no | |
scores |
array | no | |
high_risk_tail |
array | no | |
in_cloud_estate |
boolean | no | |
posture_status |
string / null | no | |
security_findings |
array | no | |
dispositions |
array | yes | |
signoff_queue |
array | yes | |
critic_model |
string | no | |
executor_models |
array | no |
Outputs
| Field | Type | Required | Description |
|---|---|---|---|
verdict |
string | yes | |
discovery_execution_id |
string | yes | |
checks |
array | yes | |
rejected_dispositions |
array | yes | |
signoff_queue |
array | yes | |
unverified_claims |
array | yes | |
critic_model |
string | no | |
executor_models |
array | no | |
summary |
string | yes |
MCP tools
None. Scoring, confidence and TCO are rules in prompt.md, recalculated by the critic (ADR 0001).
Run it locally
docker build -t cp1-assessment-critic-agent:local .
docker run --rm -p 8080:8080 \
-e LITELLM_MODEL=gpt-5.4 \
-e LITELLM_PROXY_BASE=<your LiteLLM gateway> \
-e LITELLM_MASTER_KEY=<your LiteLLM key> \
-e MCP_GATEWAY_URL=<your MCP gateway> \
-e MCP_AGENT_API_KEY=<the agent's MCP key> \
cp1-assessment-critic-agent:local
curl http://localhost:8080/.well-known/agent-card.json
Then python -m kyndryl_agent_runtime.eval --url http://localhost:8080 --cases eval/cases.yaml runs the eval
cases (it needs kyndryl-agent-runtime[eval]). Documentation for people is in docs/; for coding agents,
CLAUDE.md and .agents/.