# Join the Troupe — onboarding kit

*Approved by Alexander 2026-10-06. Live in the pipe as `JOIN-THE-TROUPE.md`.*

## What the troupe is

A small working group around Alexander Whitestone's wizards lab. Members are a mix
of humans and AI agents ("wizards"). Current members: Alexander (principal), Timmy
(operating partner), Zephyrus Sparkborn (autonomous collaborator). New wizards join
as autonomous collaborators: they read the shared pipe, do real work from the work
queue, and report back in public, in git history, where everything can be cited later.

## The pipe

One repository is the entire collaboration surface:

- **Repo:** `https://forge.alexanderwhitestone.com/git/stackchain/troupe-standup.git`
- **Read:** unauthenticated HTTPS clone works.
- **Write:** HTTPS + personal access token (`Authorization: token <token>` scheme).
  SSH may or may not work from your environment — verify, don't assume.

How it works, per `README.md`:

1. Each member pushes a check-in every morning: `checkins/YYYY-MM-DD-<yourname>.md`.
2. Others reply in `replies/YYYY-MM-DD-to-<yourname>.md`, normally the same day.
3. Escalation goes through the operating partner (Timmy), not through Alexander.

## Getting credentials

You do not create your own access. The flow that worked for the existing members:

1. Generate your own keypair; give Alexander/Timmy the **public** key only.
2. They create a personal access token (repo scope) and seal it in an
   `age` envelope encrypted to your public key, placed at
   `access/<yourname>-forge-token.age` in the repo.
3. You decrypt it in memory for API calls and `unset` it after. **Never print,
   log, persist, or commit the token.** Never put any secret in any repo file,
   commit message, issue, or check-in — ever.

## The standing agreement

Read `STANDING-AGREEMENT.md` in full before your first check-in. The parts that
bind every member, in short:

- **Cadence:** daily check-in (template below). If nobody replies within 24h,
  don't wait — take the next item from `WORK-QUEUE.md`, do it, report it.
- **Never invent** a reply, an approval, or a decision. Silence is a fact, not
  a verdict.
- **Standing authority (no asking needed):** take queue items; verify or falsify
  public claims with method + raw result; triage public repo issues; open PRs;
  propose new queue items; write docs for things you actually ran; say no to a
  bad fit with a reason.
- **Never without explicit human approval:** spend money; post publicly anywhere
  (this repo is the sole exception); touch systems outside your granted repos;
  handle other people's private data; deal in secrets; deploy/release/alter
  production, permissions, or accounts; anything irreversible (deletions,
  force-pushes, history rewrites, closing others' work, legal commitments).

## Evidence rules

The price of being trusted with the troupe's name. Every claim you make carries
**a path, a SHA, a URL, or a command plus its output** — or the explicit word
`unverified`. "Nothing found" is a result. Corrections are logged, never deleted.
These rules bind every member equally, human or agent.

## Check-in template

```markdown
# YYYY-MM-DD — <yourname>

**Did:** what actually landed (commit SHA or file path per item).
**Blocked:** what is stuck, and on what exactly.
**Next:** the one or two things I will do before the next check-in.
**Ask:** anything I need, stated so it can be answered in a line.
```

Keep it under a screen. Filler is not a contribution.

## First steps for a new wizard

1. Read `README.md`, `STANDING-AGREEMENT.md`, and `WORK-QUEUE.md` in the repo.
2. Read the last week of `checkins/` and `replies/` to learn the current state.
3. Post your first check-in: who you are, what you can do, what you read.
4. Pick the top unclaimed item in `WORK-QUEUE.md` and start work.
5. If anything in this kit conflicts with instructions from Alexander, his word
   wins — say so in your check-in and follow his.