More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Paper Writing: End-to-End Academic Manuscripts

    1

    Take a paper from idea to submission-ready draft: research memo first, an outline contract with a claims-evidence matrix, drafting in dependency order, and five review passes.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Paper Writing: End-to-End Academic Manuscripts

    Paper Writing: End-to-End Academic Manuscripts

    Example session with this skill installed

    I have experiment results for a systems paper — tables of latency numbers, an ablation half-finished, and about six weeks until the workshop deadline. I keep rewriting the introduction and never finishing. Get me from here to a submittable draft.

    • Read your context and instructions
    • Compiled the paper writing

    Diagnosis first: you are drafting out of order. The introduction keeps changing because the paper it introduces does not exist yet — so the introduction is banned until phase four of the drafting order. You start at phase one, briefly: your results exist, but the memo does not. Two to three days surveying the three threads your work touches, extracting what each claims versus shows, and logging the source of every fact you might cite. The phase ends when you can state the contribution in two sentences; if you cannot, the claim gets smaller until you can.

    Phase two, the contract, takes half a day: contribution statement in one paragraph, a one-line job per section, and the claims-evidence matrix. Your latency claims map to the finished tables — marked READY. The ablation claims map to a half-finished study — marked PLANNED with a finish date, and it is the critical path of the whole six weeks. Any claim the ablation was supposed to carry gets written into the contract's open-risks line now, so cutting it later is a decision, not a surprise.

    Phases three and four fill the remaining weeks: methods and results drafted straight from the code and logs (week 2–3), discussion and related work next, introduction and abstract in week 5 — written in days, because they now summarize a real paper. Review passes run in the final week in order: rigor (every latency number traceable to a table), coherence (introduction's promises audited against results — fix the introduction, never the numbers), citations (every reference verified, load-bearing ones spot-checked), clarity, format. Submission-ready means the scorecard reads 9+/10 with rigor and citations clean — and the ablation either finished or its claims visibly cut from the matrix.

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

    About this skill

    Papers rarely fail from bad writing — they fail from blended phases: drafting while still researching, an abstract describing the paper that was planned instead of the one that exists, citations reconstructed from memory until one of them does not exist either. This skill runs academic writing as a pipeline with strict phase separation and named deliverables at each gate. Phase one is research only: survey 3–5 literature threads, extract the gaps where the contribution lives, cross-verify disagreements, and record the source of every fact at the moment you find it — the output is a research memo, not prose, and the phase is not done until the contribution fits in two sentences. Phase two writes the outline as a contract: a one-paragraph contribution statement, a one-line job for every section, a claims-evidence matrix that marks unsupported claims before they reach the draft, and per-section page budgets, all captured in a fill-in template. Phase three drafts in dependency order — methods and results first, from the actual code and logs; discussion; related work; introduction; abstract last, as a law — with a decision rule for claims that lack evidence: run it, cite it honestly, or cut it. Phase four replaces "proofreading" with five single-concern passes (rigor, coherence, citations, clarity, format) scored on a submission-readiness scorecard that demands clean rigor and citation passes. A complete worked example on a 10-page empirical paper shows each phase's output, including the DATA_NEEDED marker that survives honestly into the draft and the coherence catch where the introduction gets fixed instead of the results. Edge cases cover negative results, multi-author merges, deadline compression, and reviewer-driven restructuring.

    • It structures and disciplines the writing process; it does not produce your results or run your experiments.
    • It is a process skill, not a typesetting tool — template mechanics for your venue are inputs, not content.
    • Citation verification still requires access to the sources; the skill defines the pass, you or your tools perform the checks.
    • It cannot make a weak contribution strong — it will, however, make the weakness visible at outline stage instead of review stage.

    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

    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 24 days ago

    Permissions

    No special permissions declared or detected

    Listed24 days ago
    Updated21 days ago

    What's inside

    Frequently Asked Questions

    No — start phases one and two now. The claims-evidence matrix exists precisely for this: finished results are marked READY, unfinished ones PLANNED or DATA_NEEDED, and the drafting order lets you write methods and the ready results while the last study runs. What you must not do is draft the introduction or abstract against results that do not exist.

    Top-to-bottom order is the failure mode the skill replaces. The introduction and abstract summarize the paper; written first, they freeze claims the results then have to be bent to match. Dependency order — methods and results, then discussion, related work, introduction, abstract — means every summary is written against something real.

    The three classic failures: a reference that does not exist (usually reconstructed from memory during drafting), a real paper cited for a claim it does not make, and correct papers with wrong attribution details. The pass verifies every citation and spot-checks the load-bearing ones against the sources — and the scorecard will not let the paper ship with that pass unclean.

    New in Research & Analysis