More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Scientific Plot Studio Pro

    1

    Editable, source-linked scientific plots and engineering figures with careful statistics and practical revisions.

    $3.99

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Scientific Plot Studio Pro

    Scientific Plot Studio Pro

    Example session with this skill installed

    Using the attributed Palmer penguins example, create a three-panel editorial figure showing flipper length versus body mass, bill-length distributions and bill depth by species/sex. Preserve all 344 source rows and original bytes. Identify panel exclusions separately, use mean ± SD only where declared, retain units and CC0 provenance, and export editable SVG, vector PDF, PNG plus captions and the analysis record. Do not invent inferential findings.

    • Read your context and instructions
    • Compiled the scientific plot studio

    Creator-local corresponding public-penguins example: 344 preserved source records. Panels A and B each use 342 records (2 excluded); panel C uses 333 (11 excluded), with explicitly labeled mean ± SD by species/sex. Generated SVG/PDF/PNG, caption, figure notes, offline report, source/project/analysis and manifest were inspected locally. This local result does not prove Agensi buyer execution or artifact retrieval. Standalone runtime delivery awaits platform confirmation.

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

    About this skill

    Create scientific and engineering figures that look deliberate and remain tied to their evidence.

    Scientific Plot Studio Pro turns supplied CSV records or explicitly identified aggregate results into editable vector figures, clear captions and a source-linked analysis report. It runs through a capable buyer agent with permitted Python/file tools; the optional browser workbench is a local helper.

    Research and technical work, covered in depth

    • Engineering benchmarks: tail latency, empirical distributions, load scaling, memory/throughput trade-offs, paired runs and multi-panel reviews.
    • ML evaluation: learning curves, seed variability, ablations, ROC, precision-recall, calibration, confusion matrices and model efficiency.
    • Experiments: distributions, paired measurements, supplied intervals and SD, SEM or independent-unit bootstrap intervals when their assumptions are confirmed.
    • Survey and observational data: proportions, ordered responses, completeness registers, cohort comparisons and complete-case correlations.
    • Time and environment: seasonal comparisons, sensor traces, log scales, event annotations and genuine gaps in incomplete series.

    51 worked projects span seven families: 48 explicitly fictional demonstrations and three attributed public-data studies. Nineteen plotting grammars, 12 visual foundations and four destination profiles (journal, engineering, presentation and editorial) support purposeful composition rather than a fixed template sequence.

    Keep numerical meaning intact

    Preserve input bytes and stable record IDs. Record delimiter/header/missing-value contracts, exclusions, per-panel counts, units and method choices. Distinguish raw observations from supplied aggregates. Do not silently deduplicate, remove outliers, fill missing readings or invent error bars, units, references, p-values or findings. Style changes do not alter numerical calculations. The helper does not certify experimental design or infer independence from row counts.

    Editing that leads to useful files

    The optional workbench offers guided column mapping and sampling-unit setup, CSV import, source/method inspection, layout and label controls, undo/redo, project save/reopen and checked export. Generated SVG retains text and vector elements; vector PDF embeds fonts; PNG provides a preview. Vector appearance editing and data-driven recomputation are different workflows, and external vector edits do not synchronize the project automatically. These are not native editable Office charts.

    Depending on host permissions, generated work can include figures, caption.md, figure-notes.json, an offline figure-report.html, original source snapshots, project JSON, analysis and output manifests. Protected Skill instructions and the example library are not exported as user work.

    Public-data examples and provenance

    Palmer penguins (Horst, Hill and Gorman; CC0) uses 344 source rows with panel-specific missing-value exclusions. UCI Iris (Fisher; CC BY 4.0) uses 150 observations; historical source discrepancies are disclosed rather than silently corrected. UCI Red Wine Quality (Cortez et al.; CC BY 4.0) uses 1,599 rows and retains original semicolon-separated source data. Attributions and dataset terms accompany these examples. Public data is descriptive and does not establish causal findings.

    Validation and compatibility disclosure

    Creator-local validation passed 59 behavioral tests and saved-output checks for all 51 projects. Actual local-browser editing, import, project reopening, downloaded output inspection, source snapshot/hash checks and an independent local rebuild were tested. Gallery images show actual creator-local outputs, with fictional and public sources identified.

    Agensi buyer-side multi-file loading, temporary execution, relative asset/font access and generated-artifact retrieval have NOT been verified. The locally tested standalone rebuild helper contains reusable creator-owned MIT runtime code; whether that code may be exported through Agensi under its protected-Skill rules is unresolved. Its delivery is not promised to Agensi buyers pending platform confirmation. Platform review is requested to verify these workflows and the boundary between generated figures/data/project files and protected Skill resources before approval. No downloadable offline Skill package or platform-independent Skill access is promised.

    Scientific expert review, external customer acceptance, physical printing and Windows/Linux GUI workflows have not been tested. Journal compliance, publication acceptance and correct conclusions are not guaranteed. Review the figure, source counts, statistical assumptions and destination requirements before use.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse All Claude Skills.

    Agensi buyer-side multi-file resource loading, temporary execution, relative assets/fonts and generated-work retrieval remain unverified; request platform review before approval. Export permission for the standalone reusable MIT rebuild runtime is unresolved and delivery is not promised. No offline Skill download/access promise. Not a general statistical-modeling, inferential-study validation or publication-certification service. No invented data, p-values or findings. External vector edits do not update source data or all formats. No native Office charts. Scientific expert review, physical printing, Windows/Linux GUI and external user acceptance remain unverified.

    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 buying a skill.

    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

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 22 days ago

    Software
    Python 3.10+NumPy 2Matplotlib 3.9+Vector editor
    Optional
    Web browser

    Matplotlib must stay on version 3; see requirements.txt. The browser is only for the optional workbench. The vector editor is for external SVG editing.

    Permissions

    Allowed Hosts

    127.0.0.1 (optional local workbench)

    File Scopes

    Only
    user-authorized
    source
    CSV/project
    files
    New
    user-selected
    output
    folders;
    preserve
    sources
    and
    previous
    versions

    No mandatory online API or seller backend. No source data upload is required by the plotting runtime. Dependency installation or extra source access must follow the host-permitted workflow.

    Listed22 days ago
    Updated14 days ago

    What's inside

    Frequently Asked Questions

    Researchers, engineers, data analysts, technical writers and students who have actual records or supplied aggregate results and need clear, revisable figures. It does not manufacture scientific results.

    51 worked projects across seven families: 48 fictional examples and three attributed public-data studies. They cover engineering, ML, experiments, surveys, time/environment, multi-panel figures and public datasets. Examples are starting points for your data.

    SVG text/vector appearance can be edited with a suitable vector editor. Project JSON and original observations support data-driven revisions through the agent/helper. Vector PDF and PNG are also generated. These are not native Office charts; external edits do not synchronize all formats.

    The workflow preserves original source snapshots and stable IDs, registers exclusions and displays panel-specific counts. SD, SEM, bootstrap and supplied intervals have distinct contracts. The user must confirm sampling units and assumptions; the helper cannot validate study design.

    No mandatory paid image service or seller-hosted backend. Actual files require a capable execution agent with permitted Python/file tools, NumPy and Matplotlib. The optional workbench uses loopback browser access.

    No offline Skill download is promised. Agensi loads protected resources through its agent connection. Generated figures/data/project files are distinct outputs. Buyer execution, retrieval and permission to export the reusable MIT rebuild runtime remain unverified and need platform confirmation before approval; that runtime delivery is not promised.

    Creator-local: 59 behavioral tests, saved outputs for all 51 projects, actual browser editing/import/download/reopen, source hashes and independent local rebuild. Agensi buyer execution, scientific expert review, physical printing and external customer acceptance remain unverified.

    Palmer penguins is CC0; UCI Iris and Red Wine Quality are CC BY 4.0 with source attribution. Preserve their credits and terms when reusing them. Your own inputs must be authorized for processing.