repo-scaffold
Skill4 starsChanged 3 months agoApache-2.0View on GitHub
Compose the initial file shape of a brand-new Horonom repository from an explicit profile, deterministically and idempotently: - Company baseline (always present): CODEOWNERS, PR template, Dependabot config, LICENSE, .gitignore/.editorconfig. - Archetype (exactly one, required): the repo's fundamental shape — see the eight below. Sets the default capability set and, just as importantly, which capabilities are **NOT_APPLICABLE** for this shape and can never be turned on. - Stack(s) (one or more, required): which language/toolchain(s) this repo composes with. A stack module…
2 other projects use a version of this.
Adopted by
Most recently taken up by
- PubCyBerry/repo-scaffoldtheir own version, first seen Nov 8, 2026
- Tamircohen28/tamirs-superpowerstheir own version, first seen Aug 3, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- Tamircohen28/tamirs-superpowers★ 2same name
- PubCyBerry/repo-scaffold★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository21 other files it gives its agents.
- .claude/CLAUDE.mdCLAUDE.md
- adr-governanceSkill
- credential-operationsSkill
- documentation-experienceSkill
- engineering-loopSkill
- governance-doctorSkill
- homebrew-tap-mergeSkill
- jira-deliverySkill
- product-validationSkill
- public-release-reconcileSkill
- release-assuranceSkill
- release-docs-syncSkill
- release-evidence-finalizeSkill
- release-qa-gateSkill
- release-security-gateSkill
- release-tag-cutSkill
- release-validate-channelsSkill
- repo-bootstrapSkill
- rust-developmentSkill
- shell-developmentSkill
- terraform-developmentSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 115
After the repo exists with its scaffolded shape, hand off to repo-bootstrap for governance adoption (skill projection, CLAUDE.md pointer …
- Distinctive: line 133
Only 3 of the 5 golden scaffolds the ticket names have real decision-logic tests: Python API service, Rust CLI, docs-only repo.
- Very specific: line 139
No integration proof that a freshly-scaffolded repo passes repo-bootstrap's own check/doctor with zero manual repair — that cross-skill …
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
<!-- horonom:generated --><!-- Source: horonomy/.github agents/skills/repo-scaffold/SKILL.md. Do not hand-edit — rerun `python3 agents/common/project_skills.py`. -->SKILL.md — repo-scaffold
Purpose
Compose the initial file shape of a brand-new Horonom repository from an explicit profile, deterministically and idempotently:
Company baseline + archetype defaults + stack module(s) + optional capabilities
= resolved scaffold- Company baseline (always present): CODEOWNERS, PR template,
Dependabot config, LICENSE,
.gitignore/.editorconfig. - Archetype (exactly one, required): the repo's fundamental shape — see the eight below. Sets the default capability set and, just as importantly, which capabilities are NOT_APPLICABLE for this shape and can never be turned on.
- Stack(s) (one or more, required): which language/toolchain(s) this
repo composes with. A stack module wires that stack's CI entrypoint and
a minimal manifest stub — it never re-explains that stack's own
build/test/lint workflow, which stays owned by the corresponding
<stack>-developmentskill (agents/skills/<stack>-development/SKILL.md, HORO-969 §1). Referencing, not duplicating. - Capabilities (optional): explicit opt-in/opt-out surfaces — Docker, Terraform CI, coverage (Codecov), SonarQube, release-binaries workflow, docs-site — layered on top of the archetype's defaults. A capability the archetype denies outright is a profile error, not a silently-omitted no-op and never a decorative empty file.
Type
Auto-used, always-applicable (no manifest.yaml — see "No manifest.yaml"
below for why). Invoke when creating a brand-new Horonom repository from an
explicit archetype/stack/capability profile — never for governance
adoption/refresh in a repo that already exists (that's repo-bootstrap;
see the boundary section immediately below).
Boundary with repo-bootstrap — read this before using either skill
repo-scaffold (this skill) creates/migrates a repo's initial shape
from an explicit profile. repo-bootstrap adopts or refreshes
Horonom governance in a repo that already exists (skill projection,
CLAUDE.md pointer block, PR template, .horonom-adoption.yaml — see
agents/skills/repo-bootstrap/SKILL.md).
They are sequential, not overlapping:
- A brand-new repo →
repo-scaffoldcomposes its initial file set from a profile. - That new repo (or any pre-existing repo lacking governance) →
repo-bootstrapadopts/refreshes company governance into it.
repo-scaffold never calls repo-bootstrap, never touches GitHub repo
settings, and never runs agents/common/project_skills.py itself — skill
projection into the new repo's .claude/skills//.codex/skills/ is
repo-bootstrap's job, applied afterward. Conversely, repo-bootstrap
never invents new-repo file shape decisions (archetype/stack/capability
composition) — that logic lives only here.
The eight archetypes
| Archetype | Shape |
library | Reusable package; coverage + SonarQube; no Docker, no release binaries. |
sdk | Client/binding package; coverage + SonarQube; no Docker, no release binaries. |
cli | Command-line tool; coverage + SonarQube + release-binaries workflow by default; Docker opt-in, not default. |
service | Long-running backend; coverage + SonarQube + Docker. |
web-app | Rendered frontend surface; coverage + SonarQube + Docker + docs-site. |
infra | Terraform-only; gets only the Terraform capability — no application CI, no coverage, no SonarQube, no Docker, no release binaries (none of those have meaning against HCL). |
docs | Documentation-only; gets only docs-site — no coverage, no SonarQube, no Docker, no release binaries, no Terraform. |
monorepo | v1: company baseline + explicitly requested capabilities only; no archetype defaults and no denials — real per-subproject composition is deferred (see Known limitations). |
Full per-surface rationale, including the "why withheld" cases, is in
references/archetype-composition.md.
When to use
- Standing up a brand-new Horonom repository and deciding what it should contain before the first commit.
- Migrating an existing repo's shape onto this profile model (e.g.
normalizing which CI/coverage/Sonar surfaces it carries) — run in
checkmode first against the existing tree.
When NOT to use
- An existing repo that already has the right shape and just needs
governance adopted/refreshed — that's
repo-bootstrap, not this skill. - Creating the actual GitHub repository or calling the GitHub API — this skill only generates file content/structure locally; repo creation is a separate, explicit step outside its scope.
- Deciding whether a stack/capability applies to an existing repo based
on evidence scanning — that's
agents/common/project_skills.py'sresolve_applicable_skills()/manifest model for skill applicability in an existing repo. A brand-new repo has no evidence yet, which is exactly whyrepo-scaffoldtakes an explicit profile instead (see manifest decision below).
Routing
- Build a profile:
{"archetype": ..., "stacks": [...], "capabilities": [...]}perscripts/repo_scaffold.py's documented schema. - Preview:
python3 scripts/repo_scaffold.py plan <profile.json>to see the resolved file list with no disk writes, orcheck <profile.json> <target-dir>to dry-run against a real target and see create/update/ collision status (exit 1 on any pending change or collision). - Apply:
python3 scripts/repo_scaffold.py write <profile.json> <target-dir>. Re-running with the same profile against the same target is a no-op (idempotent). A file that already exists and doesn't carry this generator's own provenance marker blocks the entire write — nothing is overwritten and nothing is deleted, ever, in any mode. - After the repo exists with its scaffolded shape, hand off to
repo-bootstrapfor governance adoption (skill projection, CLAUDE.md pointer block, GitHub settings).
No manifest.yaml
Unlike the stack-tagged development skills, repo-scaffold carries no
manifest.yaml. agents/common/project_skills.py's
resolve_applicable_skills() gates a skill's applicability on existing
repository evidence (Cargo.toml, pyproject.toml, etc.) — but
repo-scaffold exists precisely for repos that don't exist yet and so
cannot carry that evidence. There is no manifest shape that correctly
expresses "applies to any new repo, company-wide"; the absence of a
manifest is the correct encoding of always-applicable, the same
convention repo-bootstrap and jira-delivery already use (HORO-969 §9).
Known limitations (v1, HORO-981)
- Only 3 of the 5 golden scaffolds the ticket names have real
decision-logic tests: Python API service, Rust CLI, docs-only repo. A
TypeScript SDK with an optional Rust native binding, and an infra-only
Terraform repo's fuller shape, are not yet covered — extend
scripts/test_repo_scaffold.pybefore relying on those archetypes in anger. - No integration proof that a freshly-scaffolded repo passes
repo-bootstrap's owncheck/doctor with zero manual repair — that cross-skill proof is HORO-983's rollout scope, not this ticket's. - Plain-JSON manifest stubs (e.g.
package.json) are not emitted in v1: JSON has no comment syntax to carry this generator's provenance marker, so it can't be safely re-detected as generator-owned on a later run. Seereferences/archetype-composition.md. monorepoarchetype composition is intentionally minimal — real per-subproject archetype/stack resolution is future work.
References
references/archetype-composition.md— the full company-common-surface × archetype/stack decision table, with the withheld-surface rationale.references/reusable-ci-workflow-strategy.md— why CI is composed from small per-stack-family reusable workflow modules, not one mega-workflow.scripts/repo_scaffold.py— the generator;scripts/test_repo_scaffold.py— its test suite.examples/python-service-scaffold.md,examples/rust-cli-scaffold.md— worked invocations and their real resolved file lists.governance/engineering/agent-skill-architecture.md— HORO-969's taxonomy entry forrepo-scaffoldand the repo-bootstrap boundary.agents/skills/repo-bootstrap/SKILL.md— the sibling governance-adoption skill this one hands off to.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

