Praetor Enterprise Fleet Governance
Praetor is an open, high-integrity fleet governance and universal AI agent engineering engine for any repository fleet: you run it against your own repositories or your own fork of it. It enforces strict determinism, signed SLSA v1.0 supply chain provenance (Build Level 3 is not claimed; see releasing), bounded complexity, and cross-agent instruction synchronization. Its default owner, target framework, review bot and organisation-folder list stay empty until you configure them (framework, forge and topology settings).
Start here
Pick the path that matches why you are here:
| You want to | Read first | Then |
|---|---|---|
| Adopt Praetor in a repository | Fast adoption | Adoption and onboarding, then the HISS specification your repository is held to |
| Contribute to Praetor itself | Contributing to Praetor | Contributing, hooks and CI, then the architecture decisions |
| Operate Praetor on workstations or a cluster | Workstation install and status | Workstations, devcontainers and operations and Build, release and deployment |
Every guide is grouped under Guides. Command and MCP tool reference is in the API & CLI reference.
🏛️ Core Capabilities
| Pillar | Subsystem | Responsibility |
|---|---|---|
| Pillar I | standardsctl compile-context |
Canonical single-source-of-truth context transpiler compiling AGENTS.md into Claude, Cursor, Copilot, Windsurf, Codex, and Gemini formats (< 300 LOC budget). |
| Pillar II | standardsctl audit |
Comprehensive static analyzer sweeping source code, AST, rulesets, and baselines against the High-Integrity Systems Standard (HISS-01 through HISS-21). |
| Pillar III | standards-mcp |
Model Context Protocol server exposing multi-transport tooling (stdio, http, sse) for autonomous agents. |
| Pillar IV | standardsctl sync & plan |
Declarative repository-as-code reconciler for GitHub branch rulesets, merge policies, and label taxonomies. |
| Pillar V | internal/sentinel |
Workstation resource guardian monitoring RAM, VRAM, and disk pressure before triggering local frontier model inference. |
| Pillar VI | internal/bump |
Proactive prerelease upgrade train testing canary dependencies in isolated ephemeral git worktrees. |
🚀 Quickstart
Verify compliance and run local verification gates:
# Verify canonical agent context synchronization (HISS-16)
go run ./cmd/standardsctl compile-context --verify
# Audit repository against declared invariants and branch rulesets
go run ./cmd/standardsctl audit
# Execute full race-detected test harness
go test -v -race ./...
# Run all verification gates
make verify-all
📖 Dual-Surface AI Documentation
Praetor provides machine-readable documentation endpoints for autonomous agents:
/llms.txt: Concise, structured index of reviewed public guides and references./llms-full.txt: Source-bound compatibility map for current repository authorities; policy values remain in those sources instead of being copied into a stale second specification.