|
Generate configs, verify them, scan for issues, apply with approval. Everything sandboxed.
One platform, three tools
Run it from your terminal, let it work autonomously, or use the web interface
DojOps CLI
Your hands on the wheel
- 38 built-in DevOps skills
- 32 specialist agents
- Sandboxed execution with approval gates
- 7 LLM providers, run local with Ollama
Super Agent
Set it and forget it
- Polls Jira and GitLab for tickets
- Spawns DojOps CLI to resolve them
- Pushes code and reports on tickets
- Runs entirely on your machine
dojops-sa console
Manage licenses and billing
- License management for dojops-sa
- Billing history and payment tracking
- Offline license activation for air-gapped servers
Try dojops-sa with a free demo license
Request a 7-day demo license and let the autonomous agent resolve your tickets. No credit card required. Fill out the form and we will review your request.
Built for real infrastructure
Every step is validated, sandboxed, and logged before anything touches your repo
32 Specialist Agents
You describe the goal, DojOps picks the right agent. Terraform, Kubernetes, CI/CD, security, and more. You can also create your own agents with a single README.
Plan Before You Ship
Big goals get broken into a task graph with risk levels. You review the full plan before any file is touched. If something fails, pick up where you left off.
Zero Hallucinated YAML
LLM responses are locked to Zod schemas using provider-native JSON modes. Then validated again with terraform validate, hadolint, and kubectl --dry-run.
Every Write is Sandboxed
File writes are atomic and restricted to infrastructure paths. One command rolls everything back. Every action is recorded in a tamper-proof audit log.
10 Scanners. Automated.
Trivy, Gitleaks, Checkov, Semgrep, Hadolint, ShellCheck, npm/pip audit, SBOM generation, and license scanning. All run before configs go live. Critical findings get auto-fixed.
Build & Share Skills
38 built-in skills covering CI/CD, containers, cloud, monitoring, and security. Need something specific? Write a DOPS manifest, publish to the Hub, or grab community skills. All verified with SHA-256 checksums.
From Prompt to Production
Eleven stages between your request and a deployed config. Here's what happens.
Goal
Describe your infrastructure goal in plain English
Planner
LLM decomposes goal into tasks with risk classification
Graph
Dependency-aware topological execution graph
Executor
SafeExecutor applies policy engine and timeout checks
Generate
LLM generates configs with structured Zod output
Verify
External tool validation (terraform, hadolint, ansible)
Critic
CriticAgent reviews against best practices
Repair
Auto-fix failures and re-generate via repair loop
Execute
Sandboxed atomic writes with approval gate
Audit
Hash-chained JSONL with tamper detection
Memory
Persist execution results for future context
Built-in skills. Plug-and-play simplicity.
38 built-in DevOps skills, 7 LLM providers. Works out of the box.
DevOps Skills
LLM Providers
Bring your own model. No vendor lock-in. Run fully local with Ollama.
8 layers of defense
Enough security layers that your compliance team won't flinch at AI-generated configs
Structured output enforcement
Provider-native JSON modes so LLM output is always valid and parseable. No guessing, no fixing.
Schema validation
Every response goes through Zod safeParse(). Markdown stripping, JSON parsing, type checks. Nothing gets used without passing.
Deep verification
External tool validation: terraform validate, hadolint, kubectl --dry-run, plus built-in structure lints for GitHub Actions and GitLab CI.
Policy engine
ExecutionPolicy controls which paths are allowed, which are denied, env vars, timeouts, and file size limits. Writes are restricted to infrastructure paths only.
Approval workflows
You see a diff preview before every write. Auto-approve, auto-deny, or wire up custom callbacks for CI/CD. High-risk plans need explicit confirmation.
Sandboxed execution
Path restrictions, size limits, atomic writes via temp + rename, .bak backups, per-file audit logging. PID-based locking prevents concurrent mutations.
Immutable audit trail
Hash-chained JSONL with SHA-256 integrity verification. SIEM-compatible format. Verify tampering with a single command.
Zero telemetry
Nothing leaves your machine except requests to your chosen LLM provider. No analytics, no tracking. Run fully local with Ollama.
Get Started
$ npm i -g @dojops/cliWhat's next
Configure your LLM provider
$ dojops configInitialize in your project
$ dojops initDescribe what you need
$ dojops "Create a Terraform config for S3"Get updates on DojOps
New skills, provider integrations, and releases. Straight to your inbox. No spam, unsubscribe anytime.
Or reach us at contact@dojops.ai