lingzhi227/agent-research-skills
31 skills44.2k installs7.9k starsGitHub
Install
npx skills add https://github.com/lingzhi227/agent-research-skillsSkills in this repo
1Literature Reviewliterature-review is a Claude Code skill from lingzhi227/agent-research-skills that packages dialogue prompts extracted from STORM at Stanford. It guides agents through persona generation by selecting Wikipedia-editor-style experts, finding related topics for contextual breadth, and structuring multi-perspective discussions before writing or building. Developers reach for literature-review when starting research-heavy features, technical investigations, or content projects that benefit from structured expert viewpoints instead of a single-pass summary. The skill fits pre-implementation discovery where diverse angles reduce blind spots in requirements, architecture choices, or domain understanding.3.4kinstalls2Literature SearchThe literature-search skill. Search academic literature using Semantic Scholar, arXiv, and OpenAlex APIs. Returns structured JSONL with title, authors, year, venue, abstract, citations, and BibTeX. Use when the user needs to find papers, check related work, or build a bibliography. Expand the user's query into 2-4 complementary search queries 2. Run Semantic Scholar search (primary) with expanded queries 3. Run arXiv for very recent preprints (< 3 months) 4. Optionally run OpenAlex for broader coverage 5. Rank by: citations (0.3) + recency (0.3) + venue quality (0.2) + relevance (0.2) 7. The workflow follows the source SKILL.md contract with progressive reference loading, clear trigger phrases, and practical steps developers can apply directly in agent sessions. The workflow follows the source SKILL.md contract with progressive reference loading, clear trigger phrases, and practical steps developers can apply directly in agent sessions.2.1kinstalls3Figure Generationfigure-generation is an agent skill from lingzhi227/agent-research-skills that generate publication-quality scientific figures using matplotlib/seaborn with a three-phase pipeline (query expansion, code generation with execution, vlm visual feedback). handles bar charts, line pl. # Scientific Figure Generation Generate publication-quality figures for research papers. ## Input - `$0` — Description of the desired figure - `$1` — (Optional) Path to data file (CSV, JSON, NPY, PKL) or results directory ## Scripts ### Generate figure template ```bash python ~/.claude/skills/figure-generation/scripts/figure_template.py --type Developers invoke figure-generation during idea/research work for documentation tasks. The skill documents triggers, prerequisites, and step-by-step workflows grounded in SKILL.md. Compatible with Claude Code, Cursor, and Codex agent runtimes that load marketplace skills. Review the Security Audits panel on this listing before installing in production environments.1.7kinstalls4Latex Formattinglatex-formatting is an agent skill from lingzhi227/agent-research-skills that handle latex formatting, templates, and styling for academic papers. set up conference templates (icml, iclr, neurips, aaai, acl), fix formatting issues, manage packages, and ensure venue-specific com. # LaTeX Formatting Set up and manage LaTeX formatting for academic papers. ## Input - `$0` — Action: `setup`, `fix`, `check` - `$1` — Venue name (for `setup`) or `.tex` file path (for `fix`/`check`) ## Scripts ### Pre-submission format checker ```bash python ~/.claude/skills/latex-formatting/scripts/latex_checker.py paper/main.tex --venue neur Developers invoke latex-formatting during idea/research work for documentation tasks. The skill documents triggers, prerequisites, and step-by-step workflows grounded in SKILL.md. Compatible with Claude Code, Cursor, and Codex agent runtimes that load marketplace skills. Review the Security Audits panel on this listing before installing in production environments.1.6kinstalls5Citation Managementcitation-management is an agent skill from lingzhi227/agent-research-skills that manage bibtex citations for latex papers. harvest missing citations from a draft using semantic scholar, validate cite keys against .bib files, deduplicate entries, and format bibliography. use when w. # Citation Management Manage the full lifecycle of citations in a LaTeX paper. ## Input - `$0` — Action: `harvest`, `validate`, `add`, `format` - `$1` — Path to `.tex` or `.bib` file ## Scripts ### Validate citations (check all cite keys resolve) ```bash python ~/.claude/skills/citation-management/scripts/validate_citations.py \ --tex paper/ Developers invoke citation-management during idea/research work for documentation tasks. The skill documents triggers, prerequisites, and step-by-step workflows grounded in SKILL.md. Compatible with Claude Code, Cursor, and Codex agent runtimes that load marketplace skills. Review the Security Audits panel on this listing before installing in production environments.1.6kinstalls6Math ReasoningThe math-reasoning skill performs formal mathematical reasoning for research papers with publication-quality LaTeX output. Task types include derive for step-by-step equation derivation with boxed results, prove for direct contradiction induction or cases proofs, formalize for variable definitions and assumptions, stats for statistical test selection with p-values and effect sizes, notation for symbol tables, and verify for dimensional consistency and gradient checks. Rules require defining all symbols before use, consistent notation, numbered referenced equations, explicit assumptions, and cited lemmas. References include notation-guide.md and proof-templates.md. Use when users need mathematical derivations, theorem proofs, or statistical analysis formalization for papers.1.6kinstalls7Data AnalysisThe data-analysis skill produces rigorous statistical analysis code for research papers from CSV, JSON, pickle, or experiment log inputs plus a stated hypothesis. It structures generated Python with import, load, dataset preparation, descriptive statistics, preprocessing, analysis, and pickle export sections. A four-round review cycle checks code flaws, data handling, per-table sanity, and cross-table consistency using prompts from bundled reference files. Helper scripts stat_summary.py and format_pvalue.py recommend tests by data type, run group comparisons, and format p-values as stars or LaTeX. Allowed packages are pandas, numpy, scipy, statsmodels, sklearn, and pickle. Test selection covers t-tests, Mann-Whitney, ANOVA, chi-square, correlation, and regression variants matched to variable types. Rules require p-values on every test, confounder control, string-based column access, and never hallucinating results. Downstream skills include table-generation, figure-generation, and backward-traceability. Use when developers need reproducible analysis code and reviewed outputs for paper experiments.1.5kinstalls8Deep ResearchThe deep-research skill is designed for conduct systematic academic literature reviews in 6 phases, producing structured notes, a curated paper database, and a synthesized final report. Output is organized by phase. Installation: ~/.claude/skills/deep-research/ — scripts, references, and this skill definition. Output: .//Users/lingzhi/Code/deep-research-output/{slug}/ relative to the current working directory. Invoke when the user asks about deep research or related SKILL.md workflows.1.4kinstalls9Idea GenerationThe idea-generation skill generate novel research ideas with iterative refinement and novelty checking against literature. Score ideas on Interestingness, Feasibility, and Novelty. Use when brainstorming research directions or validating idea novelty. # Idea Generation Generate and refine novel research ideas with literature-backed novelty assessment. ## Input - `$0` - Research area, task description, or existing codebase context - `$1` - Optional: additional context (e.g., "for NeurIPS", constraints) ## Scripts ### Novelty check against Semantic Scholar ```bash python ~/.claude/skills/idea-generation/scripts/novelty_check.py \ --idea "Adaptive attention head pruning via gradient-guided importance" \ --max-rounds 5 ``` Performs iterative literature search to assess if an idea is novel. ## References - Ideation prompts (generation, reflection, novelty): `~/.claude/skills/idea-generation/references/ideation-prompts.md` ## Workflow ### Step 1: Generate Ideas Given a research area and optional code/paper context: 1. Generate 3-5 diverse research ideas 2. For each idea, provide: Name, Title, Experiment plan, and ratings 3. Use the ideation prompt templates from references ### Step.1.4kinstalls10Paper RevisionThe paper-revision skill revise papers based on reviewer feedback. Map reviewer concerns to specific sections, apply targeted edits, run additional experiments if needed, and verify improvements. Use after receiving peer review with revision requests. # Paper Revision Systematically revise papers based on reviewer feedback. ## Input - `$0` - Reviewer comments/feedback - `$1` - Current paper draft (main.tex or paper directory) ## References - Revision workflow and prompts: `~/.claude/skills/paper-revision/references/revision-prompts.md` ## Workflow ### Step 1: Parse and Prioritize Concerns For each reviewer comment: 1. Extract the specific concern 2. Classify: major revision, minor revision, question, suggestion 3. Map to affected paper section(s) 4. Prioritize: address major concerns first ### Step 2: Plan Revisions Create a revision plan: ``` Concern → Affected Section → Required Action → New Content/Experiment ``` Categories of actions: - **Clarification**: Rewrite text for clarity - **Additional experiment**: Run new experiment, add results - **New analysis**: Add ablation, statistical test, or comparison - **Structural change**: Move, merge, or split sections - **Citation*.1.4kinstalls11Paper Writing SectionThe paper-writing-section skill write a specific section of an academic paper Abstract Introduction Background Related Work Methods Experiments Results Discussion Conclusion with section-specific guidance and two-pass refinement Use when the user wants to write draft or improve a paper section Paper Section Writer Write a publication-quality section for an academic paper Input 0 Section name abstract introduction background related-work methods experimental-setup results discussion conclusion 1 Optional Path to context file research plan results prior sections Workflow Step 1 Gather Context Read the paper's existing tex files experiment logs result files and any provided context Understand title contributions methodology key results figures tables Step 2 Write the Section Load section-specific tips from references section-tips md Before every paragraph include a brief plan as a LaTeX comment Plan Step 3 Two-Pass Refinement Apply both refinement passes from references refinement-prompts md Pass 1 Fix errors unenclosed math broken refs hallucinated numbers duplicate labels Pass 2 Remove redundancies compress ensure smooth transitions References Section writing tips claude skills pap.1.4kinstalls12Experiment DesignThe experiment-design skill design experiment plans with progressive stages initial implementation baseline tuning creative research and ablation studies Plan baselines datasets hyperparameter sweeps and evaluation metrics Use when planning experiments for a research paper Experiment Design Design structured progressive experiment plans for research papers Input 0 Research idea plan or method description References 4-stage progressive experiment prompts claude skills experiment-design references stage-prompts md Scripts Generate experiment design bash python claude skills experiment-design scripts design_experiments py plan research_plan json output experiment_design json python claude skills experiment-design scripts design_experiments py method contrastive learning task classification format markdown Generates baselines ablation matrix hyperparameter grid metric selection 4-Stage Progressive Framework from AI-Scientist-v2 Stage 1 Initial Implementation Focus on getting a basic working implementation Use a simple dataset Aim for basic functional correctness Completion at least one working non-buggy implementation Stage 2 Baseline Tuning Tune hyperparameters learning rate epochs b.1.4kinstalls13Github ResearchThe github-research skill explore and analyze GitHub repositories related to a research topic Reads deep-research output discovers repos from multiple sources deeply analyzes code and produces integration blueprints GitHub Research Skill Trigger Activate this skill when the user wants to Find repos for topic GitHub research on topic Analyze open-source code for topic Find implementations of paper technique Which repos implement algorithm Uses github-research deep-research-output-dir slash command Overview This skill systematically discovers evaluates and deeply analyzes GitHub repositories related to a research topic It reads deep-research output paper database phase reports code references and produces an actionable integration blueprint for reusing open-source code Installation claude skills github-research scripts references and this skill definition Output github-research-output slug relative to the current working directory Input A deep-research output directory containing paper_db jsonl phase reports code_repos md etc 6-Phase Pipeline Phase 1 Intake Extract refs URLs keywords from deep-research output Phase 2 Discovery Multi-source broad GitHub search 50-200 repos Phase 3 Fi.1.3kinstalls14Novelty AssessmentThe novelty-assessment skill assess research idea novelty through systematic literature search. Multi-round search-evaluate loops with harsh critic persona. Binary novel/not-novel decision with justification. Use before committing to a research direction. # Novelty Assessment Rigorously assess whether a research idea is novel through systematic literature search. ## Input - `$0` - Research idea description, title, or JSON file ## Scripts ### Automated novelty check ```bash python ~/.claude/skills/idea-generation/scripts/novelty_check.py \ --idea "Your research idea description" \ --max-rounds 10 --output novelty_report.json ``` ### Literature search ```bash python ~/.claude/skills/deep-research/scripts/search_semantic_scholar.py \ --query "relevant search query" --max-results 10 ``` ## References - Assessment prompts and criteria: `~/.claude/skills/novelty-assessment/references/assessment-prompts.md` ## Workflow ### Step 1: Understand the Idea - Identify the core contribution - List the key technical components - Determine the research area and subfield ### Step 2: Multi-Round Literature Search (up to 10 rounds) For each round: 1. Generate a targeted search query 2. Search Seman.1.3kinstalls15Research PlanningThe research-planning skill design research plans and paper architectures. Given a research topic or idea, generate structured plans with methodology outlines, paper structure, dependency-ordered task lists, UML diagrams, and experiment designs. Use when starting a new research project or paper. # Research Planning Create comprehensive research plans and paper architectures from a research topic or idea. ## Input - `$0` - Research topic, idea description, or paper to reproduce ## References - Planning prompts from Paper2Code, AI-Researcher, AgentLaboratory: `~/.claude/skills/research-planning/references/planning-prompts.md` - Output schemas and templates: `~/.claude/skills/research-planning/references/output-schemas.md` ## Workflow ### Step 1: Understand the Research Context - Read any provided papers, code, or references - Identify the core research question and its significance - Assess available resources (datasets, compute, existing code) ### Step 2: Generate Research Plan Use the 4-stage planning approach (adapted from Paper2Code): 1. **Overall Plan** - Strategic overview: methodology, key experiments, evaluation metrics 2. **Architecture Design** - File structure, syst.1.3kinstalls16Survey GenerationThe survey-generation skill generate complete academic survey papers using multi-LLM parallel outline generation, RAG-based subsection writing, citation validation, and local coherence enhancement. Based on AutoSurvey pipeline. Use for writing comprehensive literature surveys. # Survey Generation Generate complete academic survey papers with structured outline, RAG-based writing, and citation validation. ## Input - `$0` - Survey topic or research area ## Scripts ### Literature search ```bash python ~/.claude/skills/deep-research/scripts/search_semantic_scholar.py \ --query "relevant search query" --max-results 50 ``` ## References - Survey prompts (outline, writing, citation, coherence): `~/.claude/skills/survey-generation/references/survey-prompts.md` ## Workflow (from AutoSurvey) ### Step 1: Collect Papers 1. Search Semantic Scholar / arXiv for papers on the topic 2. Collect 50-200 relevant papers with titles and abstracts 3. Filter by relevance and citation count ### Step 2: Generate Outline (Multi-LLM Parallel) 1. Generate N rough outlines independently (parallel) 2. Merge outlines into a single comprehensive outline 3. Expand each section into subsections 4. Edit final outl.1.3kinstalls17Code DebuggingThe code-debugging skill debug experiment code with structured error analysis. Categorize errors, apply targeted fixes with retry logic, and use reflection to prevent recurring issues. Use when experiment code fails or produces incorrect results. # Code Debugging Systematically debug experiment code with structured error categorization and fix strategies. ## Input - `$0` - Error message, stderr output, or code file with issues - `$1` - Optional: the code that produced the error ## References - Debug patterns and state machine: `~/.claude/skills/code-debugging/references/debug-patterns.md` ## Workflow ### Step 1: Categorize the Error | Category | Examples | Severity | |----------|----------|----------| | SyntaxError | Invalid syntax, indentation | Low | | ImportError | Missing module, wrong name | Low | | RuntimeError | Division by zero, shape mismatch | Medium | | TimeoutError | Infinite loop, too slow | Medium | | OutputError | Missing files, wrong format | Medium | | LogicError | Wrong results, 0% accuracy | High | ### Step 2: Analyze Root Cause 1. Read the error traceback (last 1500 chars if truncated) 2. Identify the exact line and variable causing the error 3.1.3kinstalls18Algorithm DesignThe algorithm-design skill design algorithms with LaTeX pseudocode and UML diagrams Generate algorithmic environments Mermaid class sequence diagrams and ensure consistency between pseudocode and implementation Use when formalizing methods for a paper Algorithm Design Formalize methods into algorithm pseudocode and system architecture diagrams Input 0 Method description or implementation to formalize References Algorithm and diagram templates claude skills algorithm-design references algorithm-templates md Workflow Step 1 Formalize the Algorithm 1 Define clear inputs and outputs 2 Identify the main loop recursive structure 3 Specify all parameters and their types 4 Write step-by-step pseudocode Step 2 Generate LaTeX Pseudocode Use algorithm algpseudocode environments latex begin algorithm t caption Method Name label alg method begin algorithmic 1 Require Input x parameters theta Ensure Output y State Initialize For t 1 to T State z_t gets f x_t theta If convergence criterion met State textbf break EndIf EndFor State Return y end algorithmic end algorithm Step 3 Generate UML Diagrams Mermaid Class Diagram mermaid classDiagram class Model forward x Tensor Tensor train_step batch flo.1.3kinstalls19Rebuttal WritingThe rebuttal-writing skill write point-by-point rebuttals to reviewer comments Extract concerns from reviews generate evidence-based responses and format as a structured rebuttal document Use after receiving peer review feedback Rebuttal Writing Generate structured evidence-based rebuttals to peer review comments Input 0 Reviewer comments text file or pasted directly Optional current paper draft for reference References Rebuttal prompts and format templates claude skills rebuttal-writing references rebuttal-prompts md Workflow Step 1 Parse Review Comments For each reviewer 1 Extract individual concerns questions weaknesses 2 Categorize each major concern minor concern question suggestion 3 Identify the core issue behind each concern Step 2 Generate Responses For each concern 1 Acknowledge the reviewer's point 2 Respond with evidence cite specific sections equations experiments or results from the paper 3 Describe what was done not what will be done We have added Our experiments show 4 If additional experiments are needed describe the new results concretely Step 3 Format Rebuttal Use the standard rebuttal format Response to Reviewers We thank all reviewers for their constructive1.3kinstalls20Related Work WritingThe related-work-writing skill write Related Work sections that compare and contrast prior work with your approach. Organize by theme, cite broadly, and explain how your work differs. Use when writing or improving the Related Work section of a paper. # Related Work Writing Generate publication-quality Related Work sections with proper citations and thematic organization. ## Input - `$0` - Current paper draft or method description - `$1` - Collected literature (BibTeX entries, paper summaries, or literature review notes) ## References - Related work writing prompts and strategies: `~/.claude/skills/related-work-writing/references/related-work-prompts.md` ## Workflow ### Step 1: Analyze the Paper's Contributions - Read the current paper draft (especially Methods and Introduction) - Identify the key contributions and novelty claims - List the technical components that need literature context ### Step 2: Organize Literature by Theme Group related papers into thematic clusters: - Each cluster should represent a research direction or technique - Common themes: problem formulation, methodology family, application domain, evaluation approach - Order themes from most to least relevant.1.3kinstalls21Table GenerationThe table-generation skill turns experimental result files into publication-quality LaTeX tables for research papers. It accepts comparison, ablation, descriptive, custom, and multi-dataset layouts via the results_to_table.py script with flags for bold-best, underline-second, significance stars, captions, and labels. Comparison tables list methods as rows and metrics as columns with bold best and optional mean plus std formatting. Ablation tables highlight full-model rows and can mark component presence with checkmarks. Output always uses booktabs rules, caption and label macros, textbf for winners, threeparttable notes when needed, and compact column alignment. Rules forbid hallucinated numbers: every value must match the source logs exactly. Required LaTeX packages include booktabs, multirow, multicol, and threeparttable. The skill links upstream to data-analysis and experiment-code and downstream to paper-writing and paper-compilation workflows.1.2kinstalls22Self ReviewThe self-review skill supports researchers running structured self-critique passes on drafts such as paper sections, experiment writeups, or grant paragraphs before external review. It checks claim-evidence alignment, undefined terms, missing citations, inconsistent notation, and paragraph-level clarity without rewriting entire documents blindly. Agents flag unsupported statements, suggest tighter wording, and verify that figures, tables, and text references agree. The skill is invoked after a draft exists and before paper-compilation or submission workflows in the agent-research-skills family. It emphasizes conservative edits that improve rigor while preserving author voice and existing factual content.1.2kinstalls23Slide GenerationThe slide-generation skill creates research talk slides from existing paper content, figures, and tables produced by upstream agent-research-skills. It maps one idea per slide, uses concise bullets, preserves citation markers, and highlights key results or diagrams without dumping full paper paragraphs. Agents choose layouts for title, motivation, method, results, ablation, and conclusion slides while keeping speaker-note optional detail separate. Output targets LaTeX beamer or Markdown slide decks depending on project conventions. Use when preparing conference presentations after paper sections and figures exist.1.2kinstalls24Atomic DecompositionThe atomic-decomposition skill breaks broad research questions into atomic claims, hypotheses, and experiment units that can be tested independently. It guides agents to split literature gaps, proposed methods, and evaluation criteria into small verifiable pieces before paper assembly or coding experiments. Each atom links to required evidence type such as benchmark, ablation, or user study so downstream experiment-code and paper-assembly skills inherit a traceable structure. The workflow reduces vague research plans by forcing explicit success metrics per atomic unit.1.2kinstalls25Paper CompilationThe paper-compilation skill assembles research paper artifacts into a buildable LaTeX project for submission or review. It coordinates section files, figures, tables generated by sibling skills, bibliography entries, and required packages such as booktabs and graphicx. Agents verify cross-references, labels, and citation keys, run latexmk or pdflatex build loops, and capture compile errors with targeted fixes. The skill sits downstream of paper-writing-section, table-generation, and figure-generation in the agent-research-skills pipeline. It enforces that numbers and references match upstream data sources and that wide tables use table* layouts when needed. Use when a draft paper needs final compilation, format checks, and submission-ready PDF output.1.2kinstalls26Paper AssemblyThe paper-assembly skill from agent-research-skills helps agents compile academic paper drafts by organizing atomic research notes, experiment results, and literature summaries into coherent sections. It defines workflows for mapping claims to evidence, maintaining citation placeholders, and sequencing introduction, methods, results, and discussion from prior decomposition work. Agents ensure each paragraph traces to sourced findings rather than invented metrics and cross-link to companion atomic-decomposition and experiment-code skills in the research pipeline. The skill targets researchers accelerating manuscript assembly while preserving traceability from section text back to underlying experiments.1.2kinstalls27Experiment CodeThe experiment-code skill implements runnable experiments mapped to atomic claims produced by the decomposition stage of agent-research-skills. It specifies directory layout, configuration files, logging of metrics, and reproducibility seeds so results can be cited in paper-assembly drafts. Agents scaffold training or evaluation scripts, benchmark harnesses, and ablation toggles aligned with each hypothesis rather than monolithic notebooks. Failed runs document error context for iteration while preserving traceability from code paths back to the originating atomic claim identifiers.1.2kinstalls28Paper To Codepaper-to-code is an agent research skill that converts ML research papers into complete, runnable code repositories using a 3-stage Paper2Code pipeline. Stage 1 Planning runs a four-turn conversation producing UML diagrams and a dependency graph; Stage 2 Analysis defines per-file logic; Stage 3 Coding generates files in dependency order. Input accepts a paper PDF path, paper text, or paper URL via the argument-hint parameter. Developers reach for paper-to-code when reproducing NeurIPS, ICML, or arXiv methods without manually translating equations into modules. Reference prompts live in ~/.claude/skills/paper-to-code/references/paper-to-code-prompts.md. The skill targets ML engineers and researchers who need structured, dependency-aware code generation from academic sources.1.2kinstalls29Symbolic Equationsymbolic-equation is a lingzhi227/agent-research-skills agent skill extracted from the LLM-SR codebase spanning pipeline.py, sampler.py, buffer.py, evaluator.py, and config.py. It runs a multi-island evolutionary pipeline where parallel LLM samplers propose equation function structures, an ExperienceBuffer clusters programs across islands, and evaluators score fitness against scientific inputs. Prompts ask the model to complete equation functions respecting physical meaning and variable relationships. Developers reach for symbolic-equation when fitting compact symbolic models to experimental datasets instead of opaque black-box neural regressors.1.2kinstalls30Excalidraw Skillexcalidraw-skill is an agent skill that connects coding agents to an Excalidraw canvas through 26 MCP tools for element CRUD, batch creation, duplication, layout, and canvas queries. A local Express server (default http://localhost:3000) exposes endpoints like GET /health and tools such as create_element, update_element, query_elements, and batch_create_elements. Developers reach for excalidraw-skill when they want agents to produce architecture diagrams, flowcharts, or UI mockups without manual drag-and-drop. The skill targets programmatic diagram workflows where shapes, text, arrows, and lines are created and arranged via agent tool calls rather than hand editing.1.2kinstalls31Backward Traceabilitybackward-traceability is an agent-research skill extracted from data-to-paper workflows that makes every numeric claim in a research paper traceable to code-generated output. The skill defines LaTeX patterns using \hypertarget to mark values in code output, \hyperlink to reference those values in paper text with clickable PDF jumps, and \num to evaluate derived formulas at compile time with stored explanations. Label formats connect ref_numeric_values.py, referencable_text.py, and latex_to_pdf.py style pipelines so readers click a reported metric and land on the exact artifact that produced it. Developers reach for backward-traceability when writing ML or data-science papers where reviewers demand reproducibility beyond a static results table. The skill fits teams generating LaTeX from Python pipelines who need audit-grade backward links from prose to computation.1.2kinstalls