create-slides
Use when the user asks for a slide deck, presentation, talk, or slides from a document, report, PR, notes, or topic. Builds one self-contained Machinify-branded HTML deck in a minimalist style, with headline-plus-visual slides, speaker notes, and every diagram drawn by architecture-design:create-diagram.
- License 许可
- MIT
- 最近更新
- 2026-11-14
- Stars
- ★ 1
- Forks
- 0
01 前言 Front Matter
- name
- create-slides
- description
- Use when the user asks for a slide deck, presentation, talk, or slides from a document, report, PR, notes, or topic. Builds one self-contained Machinify-branded HTML deck in a minimalist style, with headline-plus-visual slides, speaker notes, and every diagram drawn by architecture-design:create-diagram.
02 资源文件 6
- 资产文件assets/deck-template.html13.3 KB
- 资产文件assets/machinify.diagram-profile.md32.9 KB
- 参考文档evals/evals.json3.3 KB
- 参考文档references/brand.md3.4 KB
- 参考文档references/minimalism.md2.9 KB
- 脚本scripts/check_deck.py10.8 KB
03 SKILL.md 正文
Create Slides
Turn content into a minimalist Machinify HTML deck: one claim and one visual
per slide, prose in the speaker notes. The output is a single index.html
that opens from disk, navigates with the keyboard, and prints one slide per
page.
References, relative to this skill:
references/minimalism.md: the slide rules. Read it before drafting.references/brand.md: colors, type, fixed elements, icons.assets/deck-template.html: the slide catalogue to build from.assets/machinify.diagram-profile.md: the create-diagram skin for slides.scripts/check_deck.py: the rule checker.
Requirement
Diagrams come only from the architecture-design:create-diagram skill. If
it is not installed, stop and tell the user to install the
architecture-design plugin. Do not hand-draw SVG, build diagrams from HTML
boxes, or fall back to Mermaid.
Workflow
1. Intake
Read the source first: a file, a PR, a report, notes, or a topic. Then ask only for what the source does not answer:
- Audience and what they should do or believe afterwards.
- Duration. Plan about one slide per two minutes, plus title and closing.
- Ground. Light (default) or dark content slides.
- Output directory. Default: a new directory named after the deck, next to the source.
2. Storyline and outline approval
Write one claim per slide, in the order the argument needs. A claim is a sentence the audience could disagree with: "Tests own 70% of build time", not "Build time". Then pick the visual that shows the claim:
| The claim is about | Visual |
|---|---|
| Relationship, flow, structure, sequence, state, time, comparison, distribution, ranking, hierarchy | Diagram via create-diagram, which picks the type |
| One number | big-number |
| Three or four numbers | stat-strip, one marked focus |
| Three parallel concepts | icon-trio |
| One idea that needs no picture | statement slide |
There is no bullet-list slide. If a claim seems to need bullets, it is a diagram, several slides, or speaker notes.
Present the outline as a table and wait for approval before building:
| # | Layout | Eyebrow | Headline | Visual |
|---|---|---|---|---|
| 1 | title | CI review | Faster builds by Q3 | — |
| 2 | content | CI | Tests own 70% of build time | bar chart: stage durations |
Iterate on the table, not on built slides.
3. Diagrams
Prepare the profile once per deck:
- If
~/.diagram-design/profiles/machinify.mdis missing, copyassets/machinify.diagram-profile.mdthere. If it exists and differs from the bundled copy, show the difference and ask before replacing it. - Write
<deck-dir>/.diagram-designcontaining exactlyprofile: machinify. If the deck directory sits inside a git repository whose root has no marker, ask before writing one at the repository root, since it would reskin every diagram in that repository. If an existing marker names another profile, ask before changing it.
For each diagram slide, invoke the architecture-design:create-diagram skill
from the deck directory, one diagram per call, with:
- The claim the diagram must show, and the source facts behind it.
- Size preset
slide-body: the viewBox is0 0 1104 520, there is no title or eyebrow inside the SVG, and there is no footer reserve. - Profile
machinify, light or dark to match the slide. - Detail
simplifiedand audienceexecutiveunless the audience is engineers, thenbalancedandmixed. - Output
<deck-dir>/diagrams/<slug>.html, with a slug that describes the diagram.
Let create-diagram choose the diagram type and run its own checks. Then export and inline:
python3 <create-diagram-dir>/scripts/export_svg.py <deck-dir>/diagrams/<slug>.html --system-fonts
Paste the resulting <svg> element into the slide's .visual.dgm div. Drop
the <?xml ...?> line. Keep the diagram HTML next to the deck; it is the
source to edit later.
4. Assemble
Copy assets/deck-template.html to <deck-dir>/index.html. Keep the slide
blocks the outline needs, in outline order, and delete the rest. For each
slide:
- Set the eyebrow (1–3 words) and the headline (7 words or fewer, one
<strong>phrase at most). - Fill the one visual.
- Put everything else in
<aside class="notes">: explanation, sources, caveats, numbers behind the numbers. - Add class
darkto a content slide to put it on teal.
Do not add footers, page numbers, or nav dots; the template's script generates them. Do not add CSS colors; use the template's variables.
5. Check
python3 <this-skill-dir>/scripts/check_deck.py <deck-dir>/index.html
It prints JSON and exits 1 on findings. Fix every finding by cutting, moving
text to notes, or re-requesting the diagram, never by loosening a rule. Then
review what the checker cannot see, from references/minimalism.md: one
accent per slide, whitespace left empty, labels that are two to four words.
Report the deck path, the slide count, and the diagram files. Mention that
N toggles speaker notes and F toggles full screen.
Revising a deck
Edit the outline first when the change moves the argument. Re-run the diagram step only for diagrams whose claim changed, then re-export, re-inline, and re-run the checker.
04 同仓库更多 Skills
writing-design-docs
Use when the user wants to write a problem-statement design doc, says "let's plan a feature properly" / "write a design doc" / "I'm thinking about how to do X" for a non-trivial X, or before committing to a non-obvious technical direction
write-doc
Use when the user asks to write or draft a README, ADR, runbook, tutorial, how-to, explanation, or reference doc, or a caller passes "type ADR" or "type README". Picks the type, fills a template, adds diagrams. RFCs and design docs belong to writing-design-docs.
review-doc
Use when the user asks to review, critique, score, grade, or "how good is this doc" for a Markdown doc or pasted text. Read-only. Scores it on the 20 Practical Prose dimensions and emits a findings table. Does not edit; fix-doc applies the findings next.
fix-doc
Use when the user asks to fix, edit for prose quality, polish, tighten, humanize, or de-slop a Markdown doc or pasted text, or says "make this read less AI". Applies review-doc findings, generates flagged diagrams, lists substantive issues for the author, then runs the AI-tell pass last. Modifies the target.
start-session
Open a Claude Code working session in a fresh, isolated git worktree, so the user can start a task with a single /start-session <name>. Use at the START of a session when the user says "start session", "start a worktree", "spin up a worktree for ABC-123", "begin work on <ticket>", or passes a name/ticket to work under. Creates and ENTERS a worktree via the EnterWorktree tool (never `git worktree add`) under .claude/worktrees/, named exactly after the argument, on a local branch `worktree-<argument>`. Reports what the worktree inherited (`.worktreeinclude` copies, symlinked node_modules) and whether its `.env` still points at the main checkout's stack. If the argument is a JIRA key (e.g. ABC-123, PROJ-1505), also loads that story with the Atlassian CLI (`acli`) when available and summarizes the implementation ask. Also renames the session to the argument (the same effect as `/rename <NAME>`) so it's easy to find in `/resume`. Then stops and waits — it sets up the workspace and context, it does not implement.
exit-session
Tear down the ephemeral resources a Claude Code working session created, then quit the CLI — so the user can wrap up with a single /exit-session. Use when the user says "exit session", "clean up this session", "tear down everything", "wrap up and clean up resources", "remove the worktree and scratch", "stop the background processes", "shut down the stack", "clean up the agent worktrees", or asks to clean up after finishing a task — even if they only name one resource (assume they want the full sweep unless they scope it). Runs teardown permission-free: stops background processes and dev servers, Docker/compose stacks plus their volumes, networks and images, local Tilt/Kubernetes, git worktrees, merged local branches, and temp/scratch files without prompting, then exits the CLI. Covers the worktrees and Docker resources left behind by subagents spawned during the session, not just the primary agent's own worktree — both determined from git locks and Docker labels rather than from recall. Never touches resources a still-live session holds.