Runner: run a voyage remotely as the kab-automation identity #1

Open
demo-bot wants to merge 4 commits from runner-remote-voyages into main
Owner

What

The first built part of Shakedown: a runner that runs one engagement's chain (F01 Discovery → F02 Assessment → F03 Strategy) on KAB remotely, with no person's token.

  • Identity: the kab-automation client of realm kab, using client credentials. Stellwerk provisions it (crucible-platform-deploy, kab-cluster/dev component 85) and keeps the secret in crucible-kab-dev-kv. The runner reads the secret from KAB_CLIENT_SECRET only and renews the token itself.
  • Each stage:
    • checks its workflow id against the registered sub-agent workflow name;
    • builds inputs from the case and from earlier stages: execution id, approval reference and package artifact id;
    • checks them with cp1-agents' own helpers, then starts and follows the run;
    • judges the run with cp1-agents evaluate_case.py.
  • Review points (plan §5b): only in config/runner.yaml gate_projects, the runner continues on a critic pass with no failed step, and its note says the system decided. It never continues on a fail. In any other project it waits for a person and records who decided. gate_projects lists the Cymbal Bank project (5a2da83b-…, added in fb65ef9), so the runner decides review points there. Every other project waits for a person.
  • Ledger: one folder per voyage run, with --resume. No token or secret is written.
  • Also included: the ENG-CYMBAL-001 voyage, the workflow registry, a Dockerfile and an unapplied Job template.

The branch also carries two earlier plan commits that weren't on main (0890bc5: one UI in the docs portal, nightly cycle, skills; 07060db: §5b end-to-end use cases).

Tested

  • python3 -m unittest discover -s tests -t .: 15 tests against a fake KAB, all passing. They cover the whole voyage, a critic fail, waiting for a person, a person timeout, a stale workflow id, resume and dry run, plus the gate rule and token renewal.
  • Not yet run against real KAB.

Before a real run

  1. token-check: confirm the automation user can act in CP1 tenant 6ae37d…. It is an Editor of "Default Organization".
  2. run … --dry-run: the F02 and F03 workflow ids in the voyage are unverified, and a stale or individual-pattern id is refused.
  3. exec_e60e23f0a4ac is probably an individual-pattern F01 run feeding the sub-agent F02. Keep it, or run Discovery afresh.

Unconfirmed KAB details:

  • the names of finished run statuses;
  • the reviewer fields in review history;
  • _step_<n> numbering for the sub-agent workflows.

🤖 Generated with Claude Code

## What The first built part of Shakedown: a **runner** that runs one engagement's chain (F01 Discovery → F02 Assessment → F03 Strategy) on KAB remotely, with no person's token. - **Identity:** the `kab-automation` client of realm `kab`, using client credentials. Stellwerk provisions it (crucible-platform-deploy, `kab-cluster/dev` component 85) and keeps the secret in `crucible-kab-dev-kv`. The runner reads the secret from `KAB_CLIENT_SECRET` only and renews the token itself. - **Each stage:** - checks its workflow id against the registered sub-agent workflow name; - builds inputs from the case and from earlier stages: execution id, approval reference and package artifact id; - checks them with cp1-agents' own helpers, then starts and follows the run; - judges the run with cp1-agents `evaluate_case.py`. - **Review points (plan §5b):** only in `config/runner.yaml` `gate_projects`, the runner continues on a critic `pass` with no failed step, and its note says the system decided. It never continues on a `fail`. In any other project it waits for a person and records who decided. `gate_projects` lists the **Cymbal Bank** project (`5a2da83b-…`, added in `fb65ef9`), so the runner decides review points there. Every other project waits for a person. - **Ledger:** one folder per voyage run, with `--resume`. No token or secret is written. - **Also included:** the `ENG-CYMBAL-001` voyage, the workflow registry, a Dockerfile and an unapplied Job template. The branch also carries two earlier plan commits that weren't on `main` (`0890bc5`: one UI in the docs portal, nightly cycle, skills; `07060db`: §5b end-to-end use cases). ## Tested - `python3 -m unittest discover -s tests -t .`: 15 tests against a fake KAB, all passing. They cover the whole voyage, a critic fail, waiting for a person, a person timeout, a stale workflow id, resume and dry run, plus the gate rule and token renewal. - **Not yet run against real KAB.** ## Before a real run 1. `token-check`: confirm the automation user can act in CP1 tenant `6ae37d…`. It is an Editor of "Default Organization". 2. `run … --dry-run`: the F02 and F03 workflow ids in the voyage are unverified, and a stale or individual-pattern id is refused. 3. `exec_e60e23f0a4ac` is probably an individual-pattern F01 run feeding the sub-agent F02. Keep it, or run Discovery afresh. Unconfirmed KAB details: - the names of finished run statuses; - the reviewer fields in review history; - `_step_<n>` numbering for the sub-agent workflows. 🤖 Generated with [Claude Code](https://claude.com/claude-code)
demo-bot added 3 commits 2026-11-16 05:12:49 +00:00
Amended 2026-11-12 with the owner: Shakedown has no UI of its own; the KAB
Docs Portal is the one UI and Shakedown its visualisation layer for workflow
case outputs, so Shakedown serves an API that accepts the portal's realm-kab
tokens. Adds the nightly cycle (5a: harvest, generate, execute, qualify,
prove, raise pull requests automatically for next-morning review, within a
standing daily budget; never merge or deploy), skills Bosun's sub-agents work
through (5), phase 7, and the engagement-id rule from cp1-agents.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Name a customer: the system generates its use case and runs Discovery,
Assessment and Strategy remotely. It decides gates only in its own projects
and only on a critic pass, fixes failures on an agentic/<customer>-<date>
branch, builds and deploys over the shared agents (one branch at a time,
rolled back if the PR is closed), and raises one PR to main per repository.
AC8, §4, §5, §5a, §9 and §10 amended to match.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- shakedown_lib: runs F01 -> F02 -> F03 for one engagement on KAB with no
  personal token. Signs in as the kab-automation client (client credentials,
  provisioned by Stellwerk from crucible-platform-deploy) and renews the token
  itself.
- Each stage checks its workflow id against the registered sub-agent name,
  builds inputs from the case and upstream stages, starts the run, follows it
  and judges it with cp1-agents evaluate_case.py.
- Gates (plan 5b): in config/runner.yaml gate_projects only, continue on a
  critic pass with no failed step and a note saying the system decided.
  Never on a fail. Elsewhere, wait for a person and record who decided.
  gate_projects is empty for now.
- Ledger per voyage run, with --resume; nothing secret is written.
- Cymbal voyage ENG-CYMBAL-001; workflow registry; Dockerfile and an
  unapplied Job template; 15 tests against a fake KAB.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
demo-bot added 1 commit 2026-11-16 05:14:04 +00:00
The owner named the KAB project "Cymbal Bank" (5a2da83b-...; voyage
ENG-CYMBAL-001) as one of Shakedown's own. There, the runner continues a
review point only on a critic pass with no failed step (plan 5b). Every
other project still waits for a person.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This pull request can be merged automatically.
You are not authorized to merge this pull request.
View command line instructions

Checkout

From your project repository, check out a new branch and test the changes.
git fetch -u origin runner-remote-voyages:runner-remote-voyages
git checkout runner-remote-voyages
Sign in to join this conversation.
No Reviewers
No Label
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: agents/cp1-shakedown#1