SpecLoop

SpecLoop is a Rust CLI and protocol for running QA loops from product specs, business rules, acceptance criteria, and critical flows.

It gives agents a safer job:

Read specs -> inspect behavior -> collect evidence -> write findings -> suggest regression tests

Who It Is For

• AI-native product teams.
• Developers using coding agents.
• QA engineers turning product knowledge into repeatable checks.
• Technical founders who want open-core leverage without losing commercial control.

Mental Model

flowchart TD
  Specs["Specs and Rules"]
  Cache["Cold Context Cache"]
  Run["Hot Run Context"]
  Runner["Browser or Shell Runner"]
  Findings["Structured Findings"]
  Reports["Reports and Regression Suggestions"]

  Specs --> Cache
  Cache --> Runner
  Run --> Runner
  Runner --> Findings
  Findings --> Reports

Start

cargo install --path crates/specloop-cli
specloop init
specloop validate
specloop run --action "open landing page"
specloop report

specloop init creates the default .specloop/ context plus starter browser scenarios and loop runbooks. Use specloop scaffold all when you also want the optional agent, skill, command, and script assets.

Current Status

v0 is a local CLI and protocol MVP with scaffolded browser helpers. Full Chrome DevTools MCP runtime integration remains the next execution layer.