What does this skill do, and when should you use it?
pptx-author creates PowerPoint files in headless managed-agent or CMA sessions instead of editing a live PowerPoint document. It requires the deck to be written to ./out/<name>.pptx and the output directory to be created when necessary. The skill uses Python and python-pptx, supports an optional firm template, and can add standard slide content such as tables, charts, and text boxes. It is suited to file-producing automation, not sending or uploading presentations.
Writes and runs a Python script using python-pptx to create or open a presentation; uses ./templates/firm-template.pptx when provided; adds titles, tables, charts, or text boxes; saves the result as ./out/<name>.pptx; and returns the relative output path in the final response.
- A managed-agent user needs a PowerPoint file in an environment without an open Office application.
- A financial research or modeling workflow needs to package model outputs as a collectable presentation artifact.
- A team provides ./templates/firm-template.pptx and requires generated slides to follow that template.
- An automated workflow needs PNG-rendered charts embedded in PowerPoint for presentation fidelity.
How do you install this skill?
- This was a static review; script execution, python-pptx availability, template/model availability, and artifact creation were not verified.
- Sensitive financial-data handling, failure cleanup, rollback, and user-confirmation requirements are unspecified.
- There is no automated content, layout, or file-integrity quality check, so manual review is required before delivery.
- Shell / CLI
- Local filesystem
Pythonpython-pptx
The skill is located at plugins/agent-plugins/market-researcher/skills/pptx-author/SKILL.md in the anthropics/financial-services repository. In Cowork, open Settings → Plugins → Add plugin, paste https://github.com/anthropics/financial-services, and select the required agent or vertical. For Claude Code, run claude plugin marketplace add anthropics/financial-services, then claude plugin install market-researcher@claude-for-financial-services. The source does not provide a standalone installation command for this individual skill.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/anthropics/financial-services.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/plugins/agent-plugins/market-researcher/skills/pptx-author" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
In a managed-agent/CMA headless session, request a PowerPoint file and ensure it is written to ./out/<name>.pptx. Use ./templates/firm-template.pptx when available; otherwise use the default presentation layouts. The source does not define a standalone slash command or fixed trigger phrase.
What are this skill's strengths and limitations?
- Defines a precise ./out/<name>.pptx output contract for orchestration and artifact collection.
- Supports a firm template and clear one-idea-per-slide presentation conventions.
- Uses python-pptx and does not require an open Office application.
- Explicitly excludes external sending, keeping the scope clear.
- Primarily targets managed-agent/CMA headless sessions rather than live PowerPoint editing.
- Requires Python and python-pptx, but the source gives no versions or installation procedure.
- The source provides no test suite, sample output, or platform validation record.
- It does not specify detailed charting, layout behavior, or retry handling.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Headless PowerPoint Author this page ✓ Anthropic · Official | 52 · Use with care | ★ 40k | 1mo ago | Apache-2.0 |
| DeepTutor pptx Skill | 58 · Recommended | ★ 41k | 17d ago | Apache-2.0 |
| PowerPoint Template Skill Builder ✓ Anthropic · Official | 51 · Use with care | ★ 40k | 1mo ago | Apache-2.0 |
| MBB Decks | 53 · Use with care | ★ 20 | 5mo ago | MIT |
| Gorden PPT Skill | 61 · Recommended | ★ 3.6k | 4mo ago | NOASSERTION |
The source explicitly distinguishes this skill from using mcp__office__powerpoint_* to drive a live PowerPoint document; when those tools are available in Cowork plugin mode, it recommends the live-document approach.
How did FollowSkills review this skill?
The skill limits output to ./out/<name>.pptx and explicitly states that it does not email or upload files, providing useful scope and external-effect controls. It does not require user confirmation, describe sensitive-data handling, constrain script permissions, define rollback, or address dependency security, so points are deducted.
The workflow is internally coherent and covers python-pptx, template fallback, and the output path. However, it lacks dependency installation, environment checks, exception handling, diagnosable failure feedback, and key-path tests; template and model-file availability are also assumed, so points are deducted.
The target is clear: headless managed-agent/CMA sessions without an open Office application, with a defined file output and an explicit live-Office non-fit case. Inputs, trigger semantics, broader non-fit boundaries, Chinese-language support, and runtime compatibility are insufficiently specified, so points are deducted.
The document has clear sections for purpose, output contract, examples, conventions, and non-use conditions. The repository supplies an Apache-2.0 license, skill indexing, and CI validation, but the skill lacks versioning, changelog, maintenance ownership, installation notes, troubleshooting, and known-limitations disclosure, so points are deducted.
It directly addresses the core task of producing a collectible .pptx artifact and provides a basic implementation pattern and slide conventions. There is no verified representative output, complete generation workflow, quality-control step, or comparative-benefit evidence; static evidence cannot establish that only light review is needed, so points are deducted.
The contract, example, and conventions are auditable, and repository CI validates plugin structure. There are no skill-specific key-path tests, execution records, output samples, or independent corroboration, so only limited points are justified.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →