A curated marketplace of AI agents for the teams that keep production running - SRE, DevOps, DevSecOps, developer, platform. Pre-tuned for real workflows. Plug into the tools your team already runs. Bring your own LLM key. Deploy in an afternoon.
Every agent below is in production - not a demo, not a roadmap. Categorised by who typically owns the workflow, but nothing stops a DevOps team running the DevSecOps agents (and vice versa).
Alert triage that groups, deduplicates, and kills false positives before the pager fires. Incident summariser that drafts a comms update within 60 seconds of a page. Postmortem drafter from Slack + Grafana + PagerDuty context. Runbook executor that follows your existing markdown runbooks with human-in-loop checkpoints.
PagerDuty · Grafana · SigNoz · OpenObserve · Loki · Prometheus
Deploy verification against SLOs (auto-rollback on regression). Pipeline monitor that surfaces flaky stages and slow steps. Environment drift detector that catches config drift before deploys. Cost anomaly agent that flags AWS/GCP spend spikes with root-cause attribution.
GitHub Actions · Argo · Flux · Terraform · AWS Cost Explorer · GCP Billing
Vuln triage that ranks CVEs by exploitability + your actual runtime exposure (not just CVSS). Secret rotation nudges wired to the actual keys in your repos + secret stores. SBOM analysis with supply-chain risk scoring. Compliance evidence collection for SOC 2 / ISO 27001 controls, continuously.
Snyk · Trivy · Grype · Sigstore · Vault · GitHub Advanced Security · Vanta · Drata
PR review with your codebase's own conventions loaded as context. Test writing for uncovered paths (with your test framework's idioms). Dependency upgrade automation with breakage risk scoring. Error triage that clusters Sentry / Rollbar noise into actionable groups.
GitHub · GitLab · Sentry · Rollbar · your codebase (RAG-indexed)
Kubernetes resource optimiser (right-size requests/limits from actual load). Warehouse query optimiser (dbt model refactors, expensive-query hunting). Analytics agent that drafts a weekly exec digest from your BI tool. dbt review agent for model + test coverage.
Kubernetes · dbt · Snowflake · BigQuery · Metabase · Looker
Bring-your-own-LLM by default - your API key, your data path, our orchestration + agents. Nothing about your traffic lands on our infrastructure unless you deliberately opt into managed SaaS.
Anthropic Claude · OpenAI · Bedrock · self-hosted (vLLM / Ollama)
You supply the API key; we don't proxy through our servers
PagerDuty · GitHub · GitLab · Slack · MS Teams · Grafana · SigNoz · Datadog · Vault
Anthropic Agent SDK · LangGraph · Temporal · MCP servers
Trace every tool call, every retry, every human-in-loop checkpoint
Docker · Kubernetes helm chart · or managed SaaS on our infra
Managed SaaS for speed. Self-hosted in your VPC for sovereignty. Custom agent build when the catalog is close but not quite what the workflow needs.
We run the marketplace on our infrastructure. You bring your own LLM key, connect your tools via OAuth, agents run against your data - never persisted in ours. Fastest to live.
You want agents running this week and are OK with the SaaS operational shape.
Full source-available deploy into your cloud. Agents, memory, and traces all live inside your VPC. We handle setup + capability transfer. You own the ops after handover.
Data sovereignty or a regulated environment (finance, health, gov) requires everything inside your perimeter.
Not just what's in the catalog - a bespoke agent for a workflow specific to your team. Delivered as an Arcworks Engineering engagement and, if broadly useful, added to the marketplace.
You need an agent for something the catalog doesn't cover and you want it in production, not a demo.
Tell us the workflow that's costing the most human attention - the pager, the PR queue, the CVE list, the dbt refactor. A principal architect replies within one working day with a fit assessment.
Book a demo