Skip to content

Latest commit

 

History

History
140 lines (97 loc) · 4.19 KB

RELEASE_NOTES.md

File metadata and controls

140 lines (97 loc) · 4.19 KB

Agentic Operator v0.2.0 Release Notes

Release Date: April 25, 2026


What's New in v0.2.0

This release adds multi-agent examples, a pluggable workflow registry, CLI onboarding, and an in-memory cost reporter for evaluation.

These release notes describe the v0.2.0 feature set. They do not claim production readiness, full air-gap validation, durable chargeback, real OPA execution, or same-run signed evidence.


Highlights

Multi-Agent Examples

  • Research Swarm — A2A agent discovery, persona toolProfile checks, approval-threshold inputs, and a local Ollama overlay.
  • SRE Incident Response — collaborationMode: delegation, 3-tier cost-aware model routing (cheap triage → mid analysis → expensive remediation), adversarial tone on remediator, hierarchical memory. 4 alert scenarios included.

Pluggable Workflow Registry

  • @register_workflow decorator for custom LangGraph workflows
  • Auto-discovery from built-in + custom directories
  • WORKFLOW_NAME env var selects workflow at runtime
  • Ships with 3 workflows: research_swarm, code_review (fan-out security/performance/style), doc_processor (parallel entity extraction + summarization)

Agent FinOps

  • MemoryCostReporter with Prometheus metrics (agentic_workload_cost_usd, agentic_workload_tokens_total)
  • Pre-loaded pricing for OpenAI, Anthropic, Azure OpenAI, and Ollama (local = $0)
  • Budget enforcement with configurable thresholds
  • Per-workload cost annotations scraped by Prometheus

agentctl CLI Expansion

  • agentctl init — Interactive cluster onboarding wizard
  • agentctl approve — Resume PendingApproval workloads via annotation patch + Argo resume
  • agentctl workflows — List all registered workflows
  • agentctl status — Cluster health overview
  • Total: 10 commands (init, apply, get, describe, logs, cost, approve, workflows, status, version)

CRD Enhancements

  • spec.workflowName field on AgentWorkload for workflow selection
  • deepcopy regenerated for all CRD changes

🚀 Quick Start

Multi-Agent Swarm With Local Ollama

cd examples/multi-agent-swarm
cp .env.example .env        # add your API key (or use Ollama)
make up                     # starts 7 containers
make demo                   # runs full 6-stage pipeline

SRE Incident Response

cd examples/sre-incident-response
cp .env.example .env
docker compose up -d
curl -X POST http://localhost:9010/incidents \
  -H "Content-Type: application/json" \
  -d '{"alert_type":"PodCrashLoopBackOff","namespace":"default","resource":"app-server-7b9f4c6d8-x2k9p"}'

CLI Onboarding

agentctl init               # interactive cluster setup
agentctl status             # check cluster health
agentctl workflows          # list available workflows

📦 Installation

Helm

git clone https://github.com/Clawdlinux/agentic-operator-core.git
cd agentic-operator-core
helm dependency build ./charts
helm upgrade --install clawdlinux ./charts \
  --namespace agentic-system --create-namespace

agentctl

# macOS
curl -sL https://github.com/Clawdlinux/agentic-operator-core/releases/download/v0.2.0/agentctl-darwin-arm64 -o agentctl
chmod +x agentctl && sudo mv agentctl /usr/local/bin/

🔗 Links


Full Changelog

See CHANGELOG.md for the complete list of changes since v0.1.1.


🔒 Security

  • HTTPS-only enforcement for MCP endpoints
  • Webhook TLS via cert-manager, failurePolicy: Fail
  • Distroless containers, non-root user
  • Credential sanitizer in logs (OpenAI, GitHub, AWS, JWT, bearer tokens)
  • Prompt injection sanitization on scraped content
  • In-process action evaluator and separate Rego policy assets
  • See SECURITY.md for vulnerability disclosure (48h SLA)

📦 Downloads

Binaries are attached to this release. Pick the one for your platform:

  • agentctl-linux-amd64
  • agentctl-linux-arm64
  • agentctl-darwin-amd64
  • agentctl-darwin-arm64

📄 License

Apache License 2.0 — see LICENSE.