Arbora ecosystem development
For a contributor or coding agent starting with no ecosystem context, open
the public organization hub, its
build-first queue,
and contribution guide.
The queue links owner-native specs in epistemic-graph, connector SDK,
agent-utilities, arbora, WebUI, repository-manager, pipelines,
agent-terminal-ui, emerald-exchange, tunnel-manager, and the hub itself.
The canonical, complete instructions are packaged by
arbora.
Read that SKILL.md and its
bootstrap.md
from the current public checkout before changing any ecosystem repository.
The Arbora package registers the skill through its
scion_sdk.connector entry point; this catalog entry lets a
contributor discover it before installing Arbora.
Use speckit-specify, speckit-plan, speckit-tasks, speckit-analyze, sdd-implementer, and
sdd-full-lifecycle from this public skill kit for the specification and
implementation sequence. The owner repository's spec remains the authority.
Use only the repository-owned specs/<id>/ and public code/contracts as the
implementation authority. The spec must be self-contained for contributors
without private drafts, inventory, credentials, or a live deployment.
Spec-driven flow
Run this before any code. Full table and gates: the Spec-driven development section of AGENTS.md and the spec standard.
- Read
.specify/memory/constitution.md; create it from .specify/templates/constitution-template.md if absent (speckit-constitution).
- Specify the feature:
speckit-specify gathers the inputs and writes specs/<stable-id>/spec.md and requirements.md.
- Clarify: resolve every
[NEEDS CLARIFICATION: ...] marker (speckit-clarify).
- Plan:
plan.md from .specify/templates/plan-template.md, including Build contract, Production caller chain and Cross-repo seams (speckit-plan).
- Tasks:
tasks.md lines - [ ] T001 [P] [US1] <ROW-ID> ... (speckit-tasks).
- Analyze: fix the gaps and fill
checklists/requirements.md (speckit-analyze).
- Implement and land with the delivery contract in this skill (
speckit-implement); release through pipelines releases/*.yml bundles.
Never write code for a row that is not SPECIFIED. Feature directories use the stable spec ID, never a numbered prefix. The state is computed from rows, Spec: trailers and test bindings; never write status.json or a Status: line.
Rapid delivery: the 5-minute contract
Specs are already designed. The work is implementation, delivered in very small, continuously landed slices.
Time and size
- One agent delivers one PR in 5 minutes or less. No agent or sub-agent runs longer than 5 minutes; an orchestrator dispatches the next slice to a fresh agent.
- CI turns around in 5 minutes or less. PR CI runs the fast subset: lint, type check, the spec check, and only the tests and crates the diff touches. Pushes to main run the full suite.
- Never spawn sub-agents from a delivery agent.
Investigation budget
- At most 2 minutes and about 10 tool calls of reading before the first edit, and at most about 25 tool calls per PR.
- If the change is not clear by then, ship the
.1 slice (typed model plus refusal test) or skip the row with a one-line note. Do not write deferral essays.
- Trust the orchestrator's evidence and ID list; do not re-verify it.
Sizing (deterministic)
- Split a requirement when its size score is above 6, it names more than 2 code roots, or it is cross-repo. Children are
<ID>.<n>, producer first.
- Net-new work is sliced, never skipped:
.1 typed model plus validation and refusal tests, .2 the one entry point that uses it, .3+ each further behavior.
- Cross-repo moves split into one child per repo: the destination copies the behavior first, then the source switches importers and deletes.
Before every push
- Work in a real
git worktree add from origin/main. Never edit a shared checkout, never git stash, never git add -A, never force-push.
- Run the repo's own pre-commit on the changed files:
uvx --from pre-commit==4.6.0 pre-commit run --files $(git diff --name-only origin/main...HEAD). Fix every failure. No noqa, type: ignore, skip, xfail or whitelist entries to pass a gate.
- Run the targeted tests only, never a full suite.
Landing
- Every PR lands within minutes of going green; no PR sits idle. Mechanical conflicts (generated files, Markdown) are resolved automatically by the merge-train tool. Real conflicts are additive in most cases and are resolved, not deferred.
- Many open PRs land together as a merge train (one integration branch, one CI run), built with the orchestrator's merge-train tool.
- Full CI on main catches what the fast subset missed; regressions are fixed forward immediately.
- Landing in this repo:
gh pr merge --auto --merge when the PR opens. The required checks are the fast PR subset. If the PR goes DIRTY, merge origin/main into it and push.
Report: at most 8 lines: PR URL, requirement IDs, test result, and any ID:<main sha> proof for rows already on main.
Delivery practice
Follow the delivery practice reference for waiting, branches, merge holds, worktree safety and verification. The constitution holds the binding articles: XXVIII (merges and releases), XXIX (checks tell the truth), XVIII (landed means reachable), XXV (small units) and XXXII (written text is self-evident: every pull request, review or issue comment and commit message states what changed, why, the behaviour impact and how it was verified, in plain words, with no hostname, private domain, local path, internal tool, private repository, workflow jargon or private handoff pointer, and every spec row ID beside its one-line meaning).
1---2name: arbora-development3domain: development4skill_type: skill5description: >-6 Develop in any Arbora ecosystem repository from public GitHub sources.7 Route ownership, bootstrap source checkouts, and verify repository-owned8 specs across the Arbora core, public frontends, pipelines, finance and9 infrastructure agents, and the organization hub.10license: MIT11tags: [arbora, development, public-contribution]12metadata:13 source: https://skillmd.com/skills/knuckles-team/arbora-development14 version: '1.6.0'15---1617# Arbora ecosystem development1819For a contributor or coding agent starting with no ecosystem context, open20the public [organization hub](https://github.com/Knuckles-Team/.github), its21[build-first queue](https://knuckles-team.github.io/.github/build-first.html),22and [contribution guide](https://github.com/Knuckles-Team/.github/blob/main/CONTRIBUTING.md).23The queue links owner-native specs in epistemic-graph, connector SDK,24agent-utilities, arbora, WebUI, repository-manager, pipelines,25agent-terminal-ui, emerald-exchange, tunnel-manager, and the hub itself.2627The canonical, complete instructions are packaged by28[`arbora`](https://github.com/Knuckles-Team/arbora/blob/main/arbora/skills/arbora-development/SKILL.md).29Read that `SKILL.md` and its30[`bootstrap.md`](https://github.com/Knuckles-Team/arbora/blob/main/arbora/skills/arbora-development/references/bootstrap.md)31from the current public checkout before changing any ecosystem repository.32The Arbora package registers the skill through its33`scion_sdk.connector` entry point; this catalog entry lets a34contributor discover it before installing Arbora.3536Use `speckit-specify`, `speckit-plan`, `speckit-tasks`, `speckit-analyze`, `sdd-implementer`, and37`sdd-full-lifecycle` from this public skill kit for the specification and38implementation sequence. The owner repository's spec remains the authority.3940Use only the repository-owned `specs/<id>/` and public code/contracts as the41implementation authority. The spec must be self-contained for contributors42without private drafts, inventory, credentials, or a live deployment.4344## Spec-driven flow4546Run this before any code. Full table and gates: the Spec-driven development section of `AGENTS.md` and the [spec standard](https://github.com/Knuckles-Team/pipelines/blob/main/reference/spec-standard.md).47481. Read `.specify/memory/constitution.md`; create it from `.specify/templates/constitution-template.md` if absent (`speckit-constitution`).492. Specify the feature: `speckit-specify` gathers the inputs and writes `specs/<stable-id>/spec.md` and `requirements.md`.503. Clarify: resolve every `[NEEDS CLARIFICATION: ...]` marker (`speckit-clarify`).514. Plan: `plan.md` from `.specify/templates/plan-template.md`, including Build contract, Production caller chain and Cross-repo seams (`speckit-plan`).525. Tasks: `tasks.md` lines `- [ ] T001 [P] [US1] <ROW-ID> ...` (`speckit-tasks`).536. Analyze: fix the gaps and fill `checklists/requirements.md` (`speckit-analyze`).547. Implement and land with the delivery contract in this skill (`speckit-implement`); release through pipelines `releases/*.yml` bundles.5556Never write code for a row that is not SPECIFIED. Feature directories use the stable spec ID, never a numbered prefix. The state is computed from rows, `Spec:` trailers and test bindings; never write `status.json` or a `Status:` line.5758---5960## Rapid delivery: the 5-minute contract6162Specs are already designed. The work is implementation, delivered in very small, continuously landed slices.6364**Time and size**65- One agent delivers one PR in 5 minutes or less. No agent or sub-agent runs longer than 5 minutes; an orchestrator dispatches the next slice to a fresh agent.66- CI turns around in 5 minutes or less. PR CI runs the fast subset: lint, type check, the spec check, and only the tests and crates the diff touches. Pushes to main run the full suite.67- Never spawn sub-agents from a delivery agent.6869**Investigation budget**70- At most 2 minutes and about 10 tool calls of reading before the first edit, and at most about 25 tool calls per PR.71- If the change is not clear by then, ship the `.1` slice (typed model plus refusal test) or skip the row with a one-line note. Do not write deferral essays.72- Trust the orchestrator's evidence and ID list; do not re-verify it.7374**Sizing (deterministic)**75- Split a requirement when its size score is above 6, it names more than 2 code roots, or it is cross-repo. Children are `<ID>.<n>`, producer first.76- Net-new work is sliced, never skipped: `.1` typed model plus validation and refusal tests, `.2` the one entry point that uses it, `.3`+ each further behavior.77- Cross-repo moves split into one child per repo: the destination copies the behavior first, then the source switches importers and deletes.7879**Before every push**80- Work in a real `git worktree add` from `origin/main`. Never edit a shared checkout, never `git stash`, never `git add -A`, never force-push.81- Run the repo's own pre-commit on the changed files: `uvx --from pre-commit==4.6.0 pre-commit run --files $(git diff --name-only origin/main...HEAD)`. Fix every failure. No `noqa`, `type: ignore`, skip, xfail or whitelist entries to pass a gate.82- Run the targeted tests only, never a full suite.8384**Landing**85- Every PR lands within minutes of going green; no PR sits idle. Mechanical conflicts (generated files, Markdown) are resolved automatically by the merge-train tool. Real conflicts are additive in most cases and are resolved, not deferred.86- Many open PRs land together as a merge train (one integration branch, one CI run), built with the orchestrator's merge-train tool.87- Full CI on main catches what the fast subset missed; regressions are fixed forward immediately.88- Landing in this repo: `gh pr merge --auto --merge` when the PR opens. The required checks are the fast PR subset. If the PR goes DIRTY, merge `origin/main` into it and push.8990**Report**: at most 8 lines: PR URL, requirement IDs, test result, and any `ID:<main sha>` proof for rows already on main.9192## Delivery practice9394Follow the [delivery practice reference](https://github.com/Knuckles-Team/pipelines/blob/main/reference/delivery-practice.md) for waiting, branches, merge holds, worktree safety and verification. The [constitution](https://github.com/Knuckles-Team/pipelines/blob/main/.specify/memory/constitution.md) holds the binding articles: XXVIII (merges and releases), XXIX (checks tell the truth), XVIII (landed means reachable), XXV (small units) and XXXII (written text is self-evident: every pull request, review or issue comment and commit message states what changed, why, the behaviour impact and how it was verified, in plain words, with no hostname, private domain, local path, internal tool, private repository, workflow jargon or private handoff pointer, and every spec row ID beside its one-line meaning).