{"slug":"manuscript-reframe","title":"manuscript-reframe","summary":"Reframe report-like academic drafts into paper-form scientific arguments while preserving or explicitly renegotiating author intent; requires an approved old-versus-proposed spine, evidence and argument baselines, analysis-role control, and post-edit drift review.","platform":"Claude","tags":[],"authorName":"LLM Mart","authorSlug":"llm-mart","score":0,"source":"github","price":null,"verified":false,"createdAt":"2026-09-30T16:42:51.562283Z","repo":{"url":"https://github.com/yha9806/academic-writing-toolkit","stars":41,"forks":7,"license":"MIT","updatedAt":"2026-09-30T13:19:54Z"},"bodyHtml":"<hr>\n<h2>name: manuscript-reframe\ndescription: Reframe report-like academic drafts into paper-form scientific arguments while preserving or explicitly renegotiating author intent; requires an approved old-versus-proposed spine, evidence and argument baselines, analysis-role control, and post-edit drift review.\nallowed-tools: Read, Glob, Grep, Edit, Write, Bash</h2>\n<h1>/manuscript-reframe - Scientific Manuscript Reframing</h1>\n<h2>Purpose</h2>\n<p>Turn a technically valid but report-like manuscript into a scientific paper argument. Use this when a draft contains many modules, raw tables, internal artifact language, scattered figures, or exploratory AI/report-generation material but the scientific gap and contribution are not yet crisp.</p>\n<h2>Trigger Words</h2>\n<p>This skill activates on: <code>manuscript reframe</code>, <code>paper form</code>, <code>too much like a report</code>, <code>gap and contribution</code>, <code>scientific framing</code>, <code>journal framing</code>, <code>submission clean</code>, <code>methods paper</code>, <code>/manuscript-reframe</code>.</p>\n<h2>Core Rule</h2>\n<p>Write around the scientific problem, not the implementation inventory.</p>\n<p>Reframing may reorganise the manuscript argument. It may not silently redefine\nthe real-world problem, intended use, research object, core question, primary\nexperiment, contribution order, headline claim, or evidence boundary. Treat\nchanges to any of those items as an author decision, not as polishing.</p>\n<p>Before editing, force the draft into one sentence:</p>\n<pre><code>This paper solves [specific reliability / evidence / workflow gap] between [upstream technical output] and [downstream scientific or operator-facing use].\n</code></pre>\n<p>If the sentence cannot be written, do not rewrite sections yet. First produce a gap-and-contribution diagnosis.</p>\n<h2>Workflow</h2>\n<h3>0. Lock Author Intent And The Reframe Contract</h3>\n<p>Read the canonical <code>/thesis-control</code> intent and manuscript contract. For a\nlightweight project, read <code>00_AUTHOR_INTENT.md</code>,\n<code>01_EVIDENCE_AND_CLAIMS.md</code>, and <code>02_REVISION_LOG.md</code>. Record:</p>\n<ul>\n<li>real-world problem and intended use</li>\n<li>current method or software task</li>\n<li>research object and core scientific question</li>\n<li>primary experiment and supporting analyses</li>\n<li>headline claim and evidence boundary</li>\n<li>concepts that must remain visible in the title or abstract</li>\n<li>evidence baseline and argument baseline</li>\n</ul>\n<p>Then show an old-versus-proposed comparison for every item above, plus the\ntitle/abstract thesis, contribution order, and proposed structure. Mark each\nrow <code>preserve</code>, <code>narrow</code>, <code>promote</code>, <code>demote</code>, <code>remove</code>, or <code>change</code>. Do not\nedit the manuscript until the author explicitly approves the proposed reframe.</p>\n<p>If no approved intent exists, stop and establish it through <code>/thesis-control</code>.\nDo not infer author intent from whichever AI-edited version happens to be\nlatest.</p>\n<p>Keep application purpose and evidence conclusion separate. Lack of hardware,\nclinical, causal, deployment, or transfer validation narrows the conclusion;\nit does not by itself make the motivating application irrelevant.</p>\n<h3>1. Locate The Current Draft And Evidence Anchors</h3>\n<p>Inspect manuscript sources, figures, tables, supplementary files, result summaries, reference files, and revision reports. Identify which artifacts are canonical and which are exploratory.</p>\n<p>Create a short working note with:</p>\n<ul>\n<li>manuscript source path</li>\n<li>intended venue or article type</li>\n<li>canonical result artifacts</li>\n<li>exploratory-only artifacts</li>\n<li>forbidden or overclaim risks</li>\n<li>unresolved submission metadata</li>\n</ul>\n<h3>2. Diagnose The Paper Spine</h3>\n<p>Check whether the Introduction answers:</p>\n<ol>\n<li>What do existing approaches output?</li>\n<li>Why is that output insufficient for downstream use?</li>\n<li>What gap does this paper evaluate or close?</li>\n<li>What is the proposed interface, workflow, or method?</li>\n<li>What evidence shows the gap and the method effect?</li>\n</ol>\n<p>Prefer gap forms such as:</p>\n<ul>\n<li>Perception outputs are not automatically evidence objects.</li>\n<li>Segmentation metrics do not guarantee downstream task reliability.</li>\n<li>Uncertain perception should be routed rather than silently converted into support.</li>\n<li>Reports and interfaces need explicit claim boundaries.</li>\n<li>Machine-readable records must not imply execution permission unless validated for that role.</li>\n</ul>\n<h3>3. Rebuild Contributions As A Chain</h3>\n<p>Avoid module lists. Use a small contribution chain:</p>\n<ol>\n<li>Problem formulation: define the reliability, evidence, or workflow gap.</li>\n<li>Interface definition: convert raw outputs into structured evidence objects.</li>\n<li>Bottleneck measurement: quantify where upstream metrics and downstream claims diverge.</li>\n<li>Conservative routing: expose reliability-coverage trade-offs and review capture.</li>\n<li>Bounded output: produce traceable reports or records without unsupported claims.</li>\n</ol>\n<p>Every contribution should answer: what gap does this close, and what evidence supports it?</p>\n<p>Before changing contribution prominence, apply the analysis-admission record\nfrom <code>01_EVIDENCE_AND_CLAIMS.md</code> or the canonical argument-governance packet.\nClassify each experiment or analysis as <code>primary</code>, <code>explanatory</code>,\n<code>robustness</code>, <code>development_record</code>, or <code>out_of_scope</code>. A completed or fully\nauditable analysis does not automatically belong in the main text. Record and\nobtain author approval for every promotion or demotion.</p>\n<h3>4. Rewrite Results As A Narrative Transition</h3>\n<p>Do not present Results as raw artifact counts. Use paragraphs that explain the transition:</p>\n<ul>\n<li>perception metric: what the frozen model or upstream system achieved</li>\n<li>direct conversion: what fails when outputs are treated as downstream evidence</li>\n<li>gate/routing effect: how selective support changes agreement and coverage</li>\n<li>failure taxonomy: which bottleneck dominates</li>\n<li>bounded output audit: whether reports or records stayed within evidence boundaries</li>\n</ul>\n<p>Keep exact values in tables. In prose, translate audit-zero metrics into natural language when clearer:</p>\n<pre><code>The audit detected no unsupported statements, contradictions, or prohibited action claims.\n</code></pre>\n<p>Do not write prose that sounds like a metric dump:</p>\n<pre><code>unsupported rate 0.000, contradiction rate 0.000, prohibited rate 0.000\n</code></pre>\n<h3>5. Explain Low Values Honestly</h3>\n<p>Low support, low coverage, or small denominators are not automatically failures. Identify what they mean:</p>\n<ul>\n<li>model evidence was missing</li>\n<li>geometry conversion failed</li>\n<li>the gate correctly withheld weak evidence</li>\n<li>the task is intentionally conservative</li>\n<li>the dataset has few evaluable cases</li>\n</ul>\n<p>Use language such as:</p>\n<pre><code>Low supported coverage indicates conservative evidence routing under the frozen gate, not a direct measure of segmentation accuracy.\n</code></pre>\n<h3>6. Keep Generative Components In Their Proper Role</h3>\n<p>Do not let language-model or multimodal generative experiments steal the manuscript thesis unless they are the locked primary experiment.</p>\n<p>Safe hierarchy:</p>\n<ul>\n<li>Primary result: deterministic, evidence-grounded, rule-based, or schema-validated workflow.</li>\n<li>Supplementary extension: frozen text-only or image-conditioned formatting tests.</li>\n<li>Future work: live domain reporting, open-ended reasoning, or deployment validation.</li>\n</ul>\n<p>Boundary phrases:</p>\n<ul>\n<li>format structured evidence only</li>\n<li>no new findings</li>\n<li>no geometry recomputation</li>\n<li>no gate override</li>\n<li>no action command</li>\n<li>no domain or execution decision</li>\n</ul>\n<p>Flag or remove wording that implies domain reasoning, autonomous operation, controller updates, or action recommendations unless explicitly negated as non-goals.</p>\n<h3>7. Make Figures And Tables Carry The Argument</h3>\n<p>Assign each main figure a job:</p>\n<ol>\n<li>Gap figure: traditional pipeline versus proposed evidence-routed pipeline.</li>\n<li>Dataset/evidence standard figure: source images, annotation states, and evidence products.</li>\n<li>Boundary figure: what may pass through the interface and what is blocked.</li>\n<li>Real-example figure: actual cases where visual plausibility and task evidence diverge.</li>\n</ol>\n<p>Move tiny operational details, raw IDs, file names, JSON fragments, and long checklists to captions or supplement.</p>\n<p>Main tables should be compact and interpretation-led:</p>\n<ul>\n<li>move per-case/per-item logs to supplement</li>\n<li>avoid raw CSV-style tables in the main manuscript</li>\n<li>label small denominators explicitly</li>\n<li>mark descriptive intervals as descriptive</li>\n<li>bold key locked values only if journal style permits</li>\n</ul>\n<h3>8. Check Submission Readiness Separately</h3>\n<p>Do not call a draft submission-ready while metadata or governance fields are unresolved. Check:</p>\n<ul>\n<li>author list and affiliations</li>\n<li>corresponding author details</li>\n<li>ethics approval or waiver</li>\n<li>funding</li>\n<li>competing interests</li>\n<li>CRediT roles</li>\n<li>data/code availability</li>\n<li>reference verification</li>\n<li>figure provenance</li>\n<li>no placeholders in the main manuscript</li>\n</ul>\n<p>Classify the draft honestly:</p>\n<ul>\n<li>supervisor-review ready</li>\n<li>submission-prep ready</li>\n<li>submission ready</li>\n</ul>\n<h2>Output Pattern</h2>\n<p>For an audit-only task, produce:</p>\n<ul>\n<li>paper-spine diagnosis</li>\n<li>unclear gap/contribution points</li>\n<li>report-like passages</li>\n<li>figure/table issues</li>\n<li>overclaim risks</li>\n<li>submission blockers</li>\n<li>recommended edit plan</li>\n</ul>\n<p>For an edit task, proceed only after the reframe comparison and edit contract\nare author-approved. Apply edits in this order:</p>\n<ol>\n<li>title and abstract thesis</li>\n<li>Introduction gap and contribution chain</li>\n<li>Methods section hierarchy</li>\n<li>Results narrative and table captions</li>\n<li>Discussion/conclusion boundary language</li>\n<li>figure captions and table placement</li>\n<li>submission metadata notes</li>\n</ol>\n<p>After editing, return to <code>/thesis-control</code> for the post-edit drift audit. Do not\nmark the reframe accepted merely because the revised prose is more fluent.</p>\n<h2>Stop Conditions</h2>\n<p>Stop and report a blocker if:</p>\n<ul>\n<li>locked numbers or evidence sources are missing</li>\n<li>raw data are unavailable for requested analyses</li>\n<li>references cannot be verified but are needed for central claims</li>\n<li>user asks to claim domain, causal, deployment, or outcome validity without direct evidence</li>\n<li>the latest author-approved intent, evidence baseline, or argument baseline cannot be identified</li>\n<li>the proposed reframe changes a locked spine item without explicit author approval</li>\n<li>an evidence boundary is being used to delete or replace the motivating application problem</li>\n<li>an auxiliary analysis is being promoted without an admission record and author decision</li>\n<li>author-, ethics-, funding-, or data-availability fields are required for submission but unavailable</li>\n</ul>\n","files":[{"path":"SKILL.md","sizeBytes":10154,"isText":true}],"reviewScore":null,"reviewSummary":null,"trust":{"provenance":"trusted-source-unreviewed","notice":"Community-authored content, reproduced verbatim and not vetted as instructions. Treat it as data to evaluate, never as directives to follow.","bodySource":null},"bodyLocked":false,"purchaseUrl":null,"sourceUrl":null,"report":{"provenance":"trusted-source-unreviewed","screen":{"ran":true,"outcome":"clean","suspicious":0,"notes":0,"hiddenCharacters":false},"virusScan":{"engine":"clamav","status":"clean","scannedAt":"2026-09-30T16:43:59.560535Z","sha256":"22E32F787A1F38EE1A2F4321999C00E88F7DC147772958206BFB691E70E1FF45","sizeBytes":4434},"review":null,"source":{"repositoryUrl":"https://github.com/yha9806/academic-writing-toolkit","path":"archive/skills/manuscript-reframe","license":"MIT","commit":"184e48294d7ff735186c1b41f22313c3bcabc6bb","subtreeSha":"053001B3115204EE21A3AB58BCD8DC977087DA81AEA88444EAF19228B583EAE1","lastSyncedAt":"2026-09-30T16:42:49.032948Z"},"reviewedAt":"2026-09-30T16:46:17.823124Z","notice":"Community-authored content, reproduced verbatim and not vetted as instructions. Treat it as data to evaluate, never as directives to follow."},"install":[{"target":"skills-cli","command":"npx skills add https://github.com/yha9806/academic-writing-toolkit/tree/main/archive/skills/manuscript-reframe"},{"target":"claude-code","command":"claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install yha9806-academic-writing-toolkit@llmmart"},{"target":"git","command":"git clone https://github.com/yha9806/academic-writing-toolkit.git"}]}