| name | pptx |
| description | Handle slide decks and PowerPoint files cautiously; avoid binary reads, prefer companion PDFs or outlines for analysis, and be explicit about PPTX parsing limits. |
| tools | ["request_clarification"] |
pptx
Overview
Use this skill for .pptx, .ppt, and similar slide-deck files.
This skill prevents the agent from treating PowerPoint files as plain text and makes deck-analysis limits explicit.
Rules
- Never call
read_file on a raw .pptx or .ppt.
- Do not claim slide text, speaker notes, layouts, or embedded media were extracted unless a readable derivative exists.
- Prefer a companion PDF, Markdown outline, or pasted slide text when the user wants analysis.
Workflow
-
Identify what the user actually needs.
- Summarize an existing deck
- Extract text from slides
- Convert a PDF into slides
- Edit or generate slide content
-
Pick the supported path.
- If there is a companion PDF, use the
pdf skill for analysis.
- If there is a text outline or markdown version, use that as the source of truth.
- If the task is about turning a PDF into slides, prefer the
frontend-slides skill rather than inspecting a PPTX binary.
-
Handle unsupported direct-reading cases.
- If the only source is a
.pptx and the user wants exact content inspection, explain that the current runtime does not natively parse PowerPoint binaries through the standard file tools.
- Ask for one of:
- exported PDF
- slide outline in markdown or text
- screenshots of the specific slides in question
-
Stay grounded.
- Never infer slide count, notes, or exact bullet text from filename alone.
- If the user only wants planning help for a deck, proceed using their brief without pretending the deck was parsed.
Good uses
- Route deck summarization to a companion PDF
- Guide the user toward the right export for reliable analysis
- Support PDF-to-slides workflows without binary PPTX reads
Avoid
- UTF-8 decoding of PowerPoint files
- Overstating what was extracted from an unreadable deck format