Skip to content

Praetor Universal Dogfooding & Public Repo Benchmarking

Praetor's dogfooding engine ensures that governance rules are thoroughly self-verified against our own codebase and stress-tested against real-world external open-source repositories without mutating them.

🔍 Why Dogfood Against Non-Owned Public Repositories?

Dogfooding against arbitrary public repositories serves a dual purpose:

  1. For Praetor Developers:
  2. Continuously refines and stresses Praetor's AST scanner, archetype classification heuristics, and HISS invariants across diverse ecosystems (Go, Rust, Python, TypeScript, C++).
  3. Identifies edge cases in AST parsing and debt ratcheting without risking production environments.
  4. For External Developers & Organizations:
  5. Evaluate Praetor on open-source dependencies or peer codebases before adoption.
  6. Run a 100% dry-run simulation to view how Praetor would govern the repository, preview generated files, and check invariant infractions.

💻 CLI Usage (standardsctl dogfood)

1. Self-Governance Verification

Verify that your repository adheres to all HISS-16 invariants and that cross-agent context targets are synchronized:

standardsctl dogfood

2. Multi-Target Local Adoption Simulation

Simulate adoption on all repositories in your development folder in dry-run mode:

standardsctl dogfood --targets=~/dev --max-targets=15

3. Remote Non-Owned Public Repository Benchmarking

Benchmark Praetor against external public Git repositories:

# Benchmark specific public repos
standardsctl dogfood --remote=https://github.com/gin-gonic/gin,https://github.com/spf13/cobra

# Benchmark against curated popular open-source presets
standardsctl dogfood --benchmark-popular

How Remote Benchmarking Works Under the Hood:

  1. Ephemeral Shallow Clone: Praetor performs a shallow clone (git clone --depth 1 --single-branch) into an isolated temporary directory.
  2. Archetype & Invariant Scan: Executes adopt.Adopt(..., DryRun: true) and hiss.Scan(...).
  3. Readiness Grading: Computes an adoption grade (A, B, C, F) based on legacy debt and HISS invariant infractions.
  4. Instant Cleanup: Removes the temporary directory immediately upon completion. Zero disk pollution, zero repository mutation.

🤖 AI Agent MCP Interface (standards_dogfood)

AI coding agents can run dogfooding benchmarks using the standards_dogfood tool:

{
  "name": "standards_dogfood",
  "arguments": {
    "host_path": ".",
    "benchmark_popular": true
  }
}

The tool returns structured markdown reports showing context sync status, invariant audit results, and external benchmark grades.