Usage tutorial

How to use REA with an agent

REA is best understood as a repeatable investigation loop: provide an authorized target, ask focused questions, inspect evidence, and keep the result reproducible.

The four-step loop

  1. Scope: define the binary, repository, or artifact you are allowed to inspect.
  2. Observe: collect structure and metadata before making claims.
  3. Ask: give the agent one narrow question at a time.
  4. Verify: compare the answer with local evidence and record the commit and tool versions.
The online preview on this site only demonstrates the shape of the workflow. It does not execute REA or accept files.

See the fixed preview