Factory docs, home
Page navigation

Pre-release documentation. Proposed designs are labelled in the page; acceptance evidence is maintained in the build review. Local paths appear as placeholders. Current state: Build review.

Factory separates recording work, running checks and judging results. Use these explanations to understand one part. Each links to its exact contract and limits. The build review owns acceptance evidence.

Modules

Meet the modules

Distinct personalities, one recognisable Plumb family. Each portrait uses a prop to explain its module's job. Meet the twenty illustrated modules and four little record companions.

  • A wide blue bob, round spectacles and a ribbon-bound ledger. Command journal · The archivist. "Remembers the promise." Writes an accepted Command, result and Events together; a retry gets the recorded answer. Full-size PNG.
  • A squat teal bob with a compact toolbox. Local fixture worker · The tinkerer. "Small experiment, tidy bench." Runs a bounded built-in fixture and records what happened. Full-size PNG.
  • A tall violet bob comparing cards with a tuning fork. Intelligence · The matcher. "Which tool fits this task?" Records which modules may use models and why; selection is not qualification. Full-size PNG.

Module guides

  • Command journal Records each Command once, with its result and Events.

    Accepted

  • Evidence / Assurance Keeps what was observed and gives a Verdict.

    Accepted

  • Execution Records who owns each Run and Attempt, and how each ended.

    Accepted

  • Local fixture worker Runs a built-in test fixture and records what happened.

    Accepted

  • Attempt fence / workspace Gives each Attempt one gate and a private workspace.

    Accepted

  • Guarded verification / recovery Checks a fixture under one owner and settles crashed Attempts.

    Accepted

  • Confined executor Runs one small program in a macOS sandbox and reports what it saw.

    Accepted

  • Repository sealing / export and integration Seals a Candidate and moves a local branch at most once.

    Accepted

  • Portfolio Registers local projects and shows what sources report.

    Accepted

  • Dashboard A browser page on this computer for progress, usage, actions and values.

    Accepted

  • Project guidance Up to six core values per project, kept as revisions.

    Accepted

  • Required actions and blockers One list of what sources report needs attention.

    Accepted

  • Producers (Claude and OpenAI) Asks an AI model once for proposed edits; neither Claude Code nor OpenAI can run live yet.

    Accepted

  • Repository Assurance Judges a repository Candidate from stored Evidence; nothing calls it yet.

    Accepted

  • Product gate and delivery records Decides what delivery work a Product may start, and keeps each delivery record once.

    Accepted

  • Repository collector Runs a Candidate's checks in the sandbox and stores what it saw; nothing calls it yet.

    Accepted

  • Intelligence Says which modules may use an AI model, which one and why; most never do. Accepted, with its changes for the proposed discovery and post-development design; no model is qualified yet.

    Accepted

  • Lifecycle schedule Keeps the timetable of the proposed nightly checks, one named check at a time; nothing runs on a schedule yet.

    In progress

  • Discovery sources Reads the public pages you commission and keeps checkable quotes; tried only on this computer; accepted on independent review, and nothing uses it yet.

    Accepted

  • Service access Maps the outside services a Product uses and where their credentials live, by name only; increment 1, on test files.

    Accepted

  • Owner policy Records asks, answers and reported minutes; keeps stated rules without applying them.

    Accepted

  • Branch health Reads a branch's checks and workflow history and records unresolved failures; no live polling is running.

    Accepted

  • Owner digest Turns recorded facts into cited updates on the Dashboard; nothing is scheduled or sent.

    Accepted

Source: docs/explanation/modules.md