Harness AI (incl. Harness MCP Server)

Activities: Y/M	Diagnostics: P/M	Provisioning: Y/M
Event ontology: N/L	Observability: P/L	Confidence: Medium

Build style / interface — SDLC console/CLI; AI DevOps Agent; MCP server.
What it actually does — Multi-agent assist across CI/CD/security/FinOps; generates/edits pipelines and exposes platform data to external AI via MCP.
Data / telemetry — Pipelines/builds/tests/releases; FinOps.
Interoperability — MCP, APIs, VCS/CI.
Deployment model — SaaS/self-hosted components.
Notes — Focused on SDLC; pair with observability for production RCA.

UKM Snapshots:
project/run context P/M

UAM Snapshots:
audit Y/M

UDM Snapshots:
job failure hints P/M

UOM Snapshots:
N/L

UEOM Snapshots:
N/L

UPM Snapshots:
deploy/rollback Y/M

Latest updates — MCP server announced May 29, 2025.
Links — Harness AI, MCP blog, MCP GitHub