PPTX Viewing Skill — Enhanced Obsidian Notes

    1

    Turn PDF/PPTX lectures into explained Obsidian notes with selected slide visuals. Includes Python preparation and image-export helpers.

    Free

    1 useSecurity scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    PPTX Viewing Skill — Enhanced Obsidian Notes

    Example session with this skill installed

    Local helper smoke test on a synthetic one-slide deck: prepare demo lecture.pptx, then finalize a supplied study.md draft with ![[page_0001.png]]. This validates the Python helpers; it is not an Agensi buyer-side run or an AI note-quality evaluation.

    • Read your context and instructions
    • Compiled the pptx viewing

    Recorded local helper smoke test

    Environment: Linux, Python 3.12.14, PyMuPDF 1.26.6, LibreOfficeDev 26.8.0.0.alpha0 supplied by the test environment. A synthetic one-slide PPTX was used; paths included spaces and Chinese characters.

    Preparation command result (exit 0):

    Prepared 1 pages. Read slides.txt, visually inspect previews/, then draft an Obsidian Markdown note.
    

    Finalization command result (exit 0): one Markdown note and one referenced PNG were written. The final note contained exactly:

    
    # Signal decoding
    
    Acquisition records the input signal; decoding interprets it before feedback.
    
    ![[study_images/page_0001.png]]
    

    The Markdown text was supplied as a test draft; the helpers did not write or reason about its content. The preview PNG was opened and visually checked. The manifest reported one page, the final image existed, and the source file SHA-256 was unchanged. Eight separate automated checks passed, including PDF/PPTX processing, output protection and rejection of invalid references and symbolic links.

    Scope: this demonstrates local helper execution. It does not demonstrate Agensi buyer-side companion-file access or guarantee compatibility with every agent. Before using a real deck, verify script access, Python dependencies, file permissions and vision support using the README checks.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Turn lecture slides into an explanatory study note with only the visual pages that help understanding. A vision-capable AI agent reads the full deck, selects useful figures, and writes the note; the Python helpers perform file preparation and selective image export.

    Included

    • SKILL.md workflow and setup guidance.
    • PDF/PPTX preparation helper: preview PNGs, extracted text and a page manifest.
    • Referenced-image exporter: copies only cited slide previews into a sibling image folder and rewrites Markdown/Obsidian links.
    • Example note, requirements and licence notices.

    Requirements

    Python 3.9+ and PyMuPDF. PPTX conversion additionally requires LibreOffice. The host agent must support image inspection, file reads/writes, command execution, and access to both bundled Python helpers. Dependencies run in the host workspace, which may be local or cloud.

    Tested scope and compatibility

    Eight local automated checks and a separate CLI PPTX-to-note smoke test passed on Linux. Agensi accepted this ZIP and parsed its files; buyer-side MCP script delivery and execution have not been verified. Complete README capability checks before use. This is not a guarantee of compatibility with every connected agent, nor a promise of ZIP downloads or offline marketplace use.

    Limitations and privacy

    No OCR is bundled. The scripts do not perform AI reasoning, semantic image selection or automatic figure cropping. They do not delete unused previews or overwrite existing notes. They make no network requests themselves; privacy also depends on the chosen host agent and model service. AI note quality depends on the source and agent. Process only material you are authorized to use.

    Transparency

    A simpler prompt-only MIT version is free at https://github.com/lyuyao59/pptx-viewing-skill. This enhanced edition adds runnable helpers and safer file handling. Original MIT rights remain intact; dependencies retain their own licences. No affiliation with Obsidian or LibreOffice.

    Price: Free. No subscription or payment is required to access this listing.

    Agensi buyer-side delivery and execution of companion scripts remain unverified. Needs a vision-capable host with command/file tools, Python and PyMuPDF; PPTX also needs LibreOffice. No bundled OCR or script-based AI reasoning. Exports full cited slide previews, not cropped figures. Leaves unused previews intact and refuses existing outputs. Local Linux tests only; Windows/macOS not independently tested.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    1 use

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 27 days ago

    • Free with an Agensi account

    Requires

    Python 3.9+
    PyMuPDF
    LibreOffice
    Python
    macOS

    Permissions

    Terminal / Shell
    Write Files
    Read Files

    Read only the user-authorized PDF/PPTX and draft note. Write preview PNGs, extracted text, manifest, final Markdown and referenced image copies to authorized folders. No external account connection is required by the scripts. No script network requests or automatic deletions. AI host privacy settings still apply.

    Listed27 days ago

    What's inside

    Frequently Asked Questions

    The agent explains and reorganizes the deck rather than copying its text. The helpers render slide previews and copy only images cited by the agent-written draft; they do not perform AI reasoning or choose diagrams themselves.

    Windows, macOS or Linux may be used if the host can access both helpers, run Python 3.9+ with PyMuPDF, inspect images, and read/write authorized files. PPTX conversion needs LibreOffice. Local Linux tests passed; Windows/macOS and Agensi buyer-side MCP execution are not independently verified.

    The free package contains SKILL.md, two Python helpers, README setup guidance, an example note, requirements and licence notices. Access is provided through Agensi and its connected-agent workflow. Companion-script execution must be verified in the host; this listing does not promise offline use or a buyer ZIP download.

    The scripts themselves make no network requests. Files are processed in the host workspace; a cloud agent may send source content or images to its model service according to its own settings. Use only authorized material and review your host privacy settings.

    The AI agent selects useful slide pages and writes the draft. The finalizer copies only referenced full-slide PNGs, rewrites links into a sibling image folder, and leaves the other previews intact. It does not automatically crop figures or delete files.

    Preparation can render scanned PDFs as preview images, but extracted text may be empty. A vision-capable agent must inspect those images; no OCR engine is bundled. State visibility limitations instead of inventing contents.

    New in Productivity