Prosperous documentation

Remote management for anything that runs Orbis software - a prepared machine, or orbistoun. Register it, ask what it can currently do, put a payload on it, read its log, move files, watch its output.

New here? The root README has the pitch. Then, depending on what you came for:

The words

• GLOSSARY.md - the services on the target, chains against health checks, scan roots, portable mode and Porthole. The collection's glossary covers standard ELF and the words that mean something else in the sibling repositories.

Guide

• features/README.md - Features & screens index mapping GUI tabs to CLI commands.
• features/user-guide.md - Paths, portable mode, network daemons, and zero telemetry.
• features/targets.md - Target registration and health check matrix.
• features/logs.md - Live kernel telemetry streamer.
• features/files.md - Remote storage browser and staging.
• features/titles.md - Title supervisor and process lifecycle.
• features/shell.md - Remote command shell console.
• features/payloads.md - Sending payloads and memory injection.
• features/library.md - Titles, saves, packages, and storage.
• Getting started - from a download to a target that answers.
• BUILDING.md - bin/prosperous, what each verb does, what check runs and in what order, and what CI runs. You do not need this to use Prosperous; the releases page has binaries.

Reference

• USAGE.md - what pros reports and how to drive it, for somebody already running the tool. This was the middle of the README.
• DESIGN.md - what the tool is and why it is shaped this way: one instrument, two transports, and the library underneath both.
• CAPABILITIES.md - a design note on being worth running once every payload it currently names has been superseded. Steps 1 and 2 are built; the rest is design.
• VIDEO.md - watching and diffing: two problems that look like one and share no code.
• PAD.md - a proposal, not a decision, about where the pad should live. Nothing in it is built.

Project memory

• DECISIONS.md - a generated index over decisions/, one file per entry. Every non-obvious choice, numbered, with the reasoning, including the ones reversed on evidence.
• WORKLOG.md - what was done, in order, with the surprises.
• ROADMAP.md - what is wrong, what is built and not reachable, what is missing, and what is next.

These are dated records. What they say was true when it was written and is not corrected afterwards. Correcting a log is falsifying it.

Shared rules - provenance, naming, decision logs, gates - are in the OOPS conventions and not restated here.

Adding to a log

The decisions log is a directory with a generated index. Add a file under decisions/, then regenerate the table:

tools/split-decisions.sh --index prosperous

Do not edit the index by hand - it is overwritten. The split exists because two sessions appending to one file collide, which is where the duplicate numbers and out-of-order entries came from, and because a log past half a megabyte stops rendering on GitHub entirely.