data2dsl

04-koru-remediation-feed

Demonstrates generating machine-actionable new-project.remediation-intent/v1 manifests consumed by semcod/koru for closed-loop self-healing (DETECT → PLAN → EXECUTE → VERIFY → HEAL).

Fixtures

• bundle.json: A comparison bundle containing discrepancy data.
• expected-remediation-intent.json: Structured remediation plan with action items (synchronize_metric), status PROPOSED, and pinned evidence hashes.

Running via CLI and Python API

# Via CLI
python src/data2dsl_cli.py feed-koru -b bundle.json -o remediation_intent.json

# Via Python API
from data2dsl_remediation import format_remediation_intent

intent = format_remediation_intent(bundle)
assert intent["status"] == "PROPOSED"
assert len(intent["actions"]) == 1
assert intent["actions"][0]["action_type"] == "synchronize_metric"