Repo Prep
Interactively scaffold the legal, metadata, community, and compliance files a
repository needs before publication — consistently, from a central author
profile, with author choices respected per repo.
When to use
- Preparing a new or existing repo for GitHub / open-sourcing.
- Adding missing LICENSE, NOTICE, AUTHORSHIP, README sections, or package metadata.
- Writing a defensible AUTHORSHIP record for AI-assisted work (the wizard).
- Adding community docs or
.github/ templates.
- Checking EU/Germany compliance obligations for a repo.
Operating principles
- Interactive, not assumed. Ask the user for choices (see Asking questions
below); respect the author profile's
defaults but confirm per repo. Don't
blindly dump templates.
- Idempotent. Detect what already exists; skip or refresh, never clobber
silently. Show a plan before writing.
- Evidence over assertion (especially authorship): mine the repo for facts.
- Confirm outward-facing actions (anything via
gh, pushes) before running.
- Follow the user's git conventions — commit only when asked.
Asking questions (interview transport)
Prefer cenno when it is installed — it gives a nicer panel, progress dots,
and voice input (great for the reflective authorship answers).
- Detect cenno once per run:
ToolSearch for
mcp__cenno__ask_sequence. If the tools resolve, use cenno; otherwise fall
back to AskUserQuestion (choices) / plain conversational prompts (free text).
- When using cenno:
- Multi-question flows (the authorship wizard, scope selection) →
mcp__cenno__ask_sequence with flow: "question"; it auto-fills progress dots.
- Single questions →
mcp__cenno__ask_user.
- Use
input.kind: "choice" + choices for pick-one; "confirm" for yes/no;
"voice_text" for reflective free-text (Decisions, Judgment, Direction, Art
Direction); "scale" for ratings.
- Parse the returned JSON (
{answers: [...]} / {answer, ...}). Handle
{answered: false} (timeout) by retrying or falling back.
- If cenno is wanted but not installed, mention it can be set up via the
cenno skill; do not block — fall back gracefully.
Central author profile (single source of truth)
Identity, promo links ("ads"), and defaults live in
~/.config/personal-os/repo-prep/profile.toml (override with REPO_PREP_PROFILE).
- On first run, if it doesn't exist, copy
assets/templates/profile.example.toml
there and walk the user through filling it (name, email, github, url, promo
projects, defaults). It is pre-seeded for this author.
- Read it at the start of every run to resolve
{{author}}, {{email}},
{{github}}, the promo block, and default choices.
Workflow
- Detect — inspect the repo: manifest/ecosystem (
pyproject.toml,
package.json, Cargo.toml, go.mod), existing files, git remote, default
branch, whether it's already on GitHub. Read the author profile.
- Scope — confirm which components to do via
AskUserQuestion (core /
community / .github / promo / compliance / gh remote). Pre-fill from profile
defaults.
- Resolve placeholders — once, per
references/checklist.md.
- Generate — for each chosen component, create files from
assets/templates/ and assets/github/, filling placeholders. Use
scripts/fetch_license.py for LICENSE. Edit package metadata inline.
- Authorship — if doing AUTHORSHIP.md, run the wizard (below), don't just
fill the template.
- Compliance — if requested or the author is EU-based, run the gating
questions in
references/eu-germany-compliance.md and add only what applies.
- gh remote (optional) — confirm, then create/configure the repo.
- Verify — run the test suite (metadata edits can break builds); show the
diff; commit only if asked.
Full item-by-item detail, placeholder table, and ecosystem detection:
references/checklist.md (read it before generating).
The authorship wizard
A first-class feature. To produce a defensible AUTHORSHIP.md for AI-assisted
work, read and follow references/authorship-wizard.md. It:
- gathers external evidence first (ADRs, git log, changelog, AI session logs,
CLAUDE.md/AGENTS.md, design tokens) and offers a draft to edit;
- asks the jurisdiction first, then interviews across five sections —
Decisions, Exercise of Judgment, Goal Setting & Direction, Art Direction,
AI Implementation;
- writes a Legal & Copyright section with jurisdiction-aware framing
(US human-authorship bar, EU/Germany "own intellectual creation", UK CGW);
- always states it is not legal advice.
Output template: assets/templates/AUTHORSHIP.md.
EU / Germany compliance
references/eu-germany-compliance.md covers CRA (SBOM, vuln reporting), AI Act
Art. 50 transparency, German Impressum (DDG), GDPR, the Product Liability
Directive, and the EAA. Obligations are conditional — ask the gating
questions and add only what applies. Not legal advice.
Resources
assets/licenses/ — bundled MIT, BSD-3-Clause (offline); others fetched.
assets/templates/ — AUTHORSHIP, CODE_OF_CONDUCT, CONTRIBUTING, SECURITY,
NOTICE, CHANGELOG, profile.example.toml.
assets/github/ — PR + issue templates, dependabot, CI workflows (python/node).
scripts/fetch_license.py — write a filled LICENSE for any SPDX id.
references/ — checklist.md, licenses.md, authorship-wizard.md,
eu-germany-compliance.md.
1---
2name: repo-prep
3description: Interactively prepare a code repository for publication — LICENSE, NOTICE, AUTHORSHIP, README sections, package metadata, .gitignore, community docs (CONTRIBUTING/CODE_OF_CONDUCT/SECURITY/CHANGELOG), .github templates (issues/PR/CI/dependabot), a promo "more from the author" block, and conditional EU/Germany legal compliance (CRA, AI Act, GDPR, Impressum, product liability). Includes a dedicated authorship wizard that documents human decisions, judgment, direction, and art direction with jurisdiction-aware legal framing. Use when setting up a new repo for release, adding missing legal/meta/community files, writing an AUTHORSHIP record for AI-assisted work, or asking "prepare this repo for GitHub / open-sourcing".
4metadata:
5 source: https://skillmd.com/skills/glebis/repo-prep
6---
7
8# Repo Prep
9
10Interactively scaffold the legal, metadata, community, and compliance files a
11repository needs before publication — consistently, from a central author
12profile, with author choices respected per repo.
13
14## When to use
15
16- Preparing a new or existing repo for GitHub / open-sourcing.
17- Adding missing LICENSE, NOTICE, AUTHORSHIP, README sections, or package metadata.
18- Writing a defensible **AUTHORSHIP** record for AI-assisted work (the wizard).
19- Adding community docs or `.github/` templates.
20- Checking **EU/Germany compliance** obligations for a repo.
21
22## Operating principles
23
24- **Interactive, not assumed.** Ask the user for choices (see *Asking questions*
25 below); respect the author profile's `defaults` but confirm per repo. Don't
26 blindly dump templates.
27- **Idempotent.** Detect what already exists; skip or refresh, never clobber
28 silently. Show a plan before writing.
29- **Evidence over assertion** (especially authorship): mine the repo for facts.
30- **Confirm outward-facing actions** (anything via `gh`, pushes) before running.
31- Follow the user's git conventions — commit only when asked.
32
33## Asking questions (interview transport)
34
35Prefer **cenno** when it is installed — it gives a nicer panel, progress dots,
36and voice input (great for the reflective authorship answers).
37
381. **Detect cenno once per run:** `ToolSearch` for
39 `mcp__cenno__ask_sequence`. If the tools resolve, use cenno; otherwise fall
40 back to `AskUserQuestion` (choices) / plain conversational prompts (free text).
412. **When using cenno:**
42 - Multi-question flows (the authorship wizard, scope selection) →
43 `mcp__cenno__ask_sequence` with `flow: "question"`; it auto-fills progress dots.
44 - Single questions → `mcp__cenno__ask_user`.
45 - Use `input.kind: "choice"` + `choices` for pick-one; `"confirm"` for yes/no;
46 `"voice_text"` for reflective free-text (Decisions, Judgment, Direction, Art
47 Direction); `"scale"` for ratings.
48 - Parse the returned JSON (`{answers: [...]}` / `{answer, ...}`). Handle
49 `{answered: false}` (timeout) by retrying or falling back.
503. **If cenno is wanted but not installed**, mention it can be set up via the
51 `cenno` skill; do not block — fall back gracefully.
52
53## Central author profile (single source of truth)
54
55Identity, promo links ("ads"), and defaults live in
56`~/.config/personal-os/repo-prep/profile.toml` (override with `REPO_PREP_PROFILE`).
57
581. On first run, if it doesn't exist, copy `assets/templates/profile.example.toml`
59 there and walk the user through filling it (name, email, github, url, promo
60 projects, defaults). It is pre-seeded for this author.
612. Read it at the start of every run to resolve `{{author}}`, `{{email}}`,
62 `{{github}}`, the promo block, and default choices.
63
64## Workflow
65
661. **Detect** — inspect the repo: manifest/ecosystem (`pyproject.toml`,
67 `package.json`, `Cargo.toml`, `go.mod`), existing files, `git remote`, default
68 branch, whether it's already on GitHub. Read the author profile.
692. **Scope** — confirm which components to do via `AskUserQuestion` (core /
70 community / .github / promo / compliance / gh remote). Pre-fill from profile
71 `defaults`.
723. **Resolve placeholders** — once, per `references/checklist.md`.
734. **Generate** — for each chosen component, create files from
74 `assets/templates/` and `assets/github/`, filling placeholders. Use
75 `scripts/fetch_license.py` for LICENSE. Edit package metadata inline.
765. **Authorship** — if doing AUTHORSHIP.md, run the wizard (below), don't just
77 fill the template.
786. **Compliance** — if requested or the author is EU-based, run the gating
79 questions in `references/eu-germany-compliance.md` and add only what applies.
807. **gh remote** (optional) — confirm, then create/configure the repo.
818. **Verify** — run the test suite (metadata edits can break builds); show the
82 diff; commit only if asked.
83
84Full item-by-item detail, placeholder table, and ecosystem detection:
85**`references/checklist.md`** (read it before generating).
86
87## The authorship wizard
88
89A first-class feature. To produce a defensible `AUTHORSHIP.md` for AI-assisted
90work, **read and follow `references/authorship-wizard.md`**. It:
91
92- gathers external evidence first (ADRs, git log, changelog, AI session logs,
93 `CLAUDE.md`/`AGENTS.md`, design tokens) and offers a draft to edit;
94- asks the **jurisdiction** first, then interviews across five sections —
95 **Decisions, Exercise of Judgment, Goal Setting & Direction, Art Direction,
96 AI Implementation**;
97- writes a **Legal & Copyright** section with jurisdiction-aware framing
98 (US human-authorship bar, EU/Germany "own intellectual creation", UK CGW);
99- always states it is **not legal advice**.
100
101Output template: `assets/templates/AUTHORSHIP.md`.
102
103## EU / Germany compliance
104
105`references/eu-germany-compliance.md` covers CRA (SBOM, vuln reporting), AI Act
106Art. 50 transparency, German Impressum (DDG), GDPR, the Product Liability
107Directive, and the EAA. Obligations are **conditional** — ask the gating
108questions and add only what applies. Not legal advice.
109
110## Resources
111
112- `assets/licenses/` — bundled MIT, BSD-3-Clause (offline); others fetched.
113- `assets/templates/` — AUTHORSHIP, CODE_OF_CONDUCT, CONTRIBUTING, SECURITY,
114 NOTICE, CHANGELOG, profile.example.toml.
115- `assets/github/` — PR + issue templates, dependabot, CI workflows (python/node).
116- `scripts/fetch_license.py` — write a filled LICENSE for any SPDX id.
117- `references/` — `checklist.md`, `licenses.md`, `authorship-wizard.md`,
118 `eu-germany-compliance.md`.