From research-workspace
Extracts claims, figures, and evidence from academic paper drafts (Word/LaTeX/MD/Obsidian), Zotero metadata into .paper/claims.yml and figures.yml for AI-assisted writing/revision.
npx claudepluginhub wenyuchiou/ai-research-skills --plugin research-workspaceThis skill uses the workspace's default tool permissions.
Bridge between research-hub and `academic-writing-skills`. Reads a
Orchestrates multi-agent writing of academic papers or proposals from research artifacts, with evidence-grounded prose, MAGI cross-review, and quality validation.
Manages project-scoped literature reviews in Obsidian: organizes papers in Sources/Papers, synthesizes insights in Knowledge notes like Literature Overview and Research Gaps, drafts in Writing, maps to literature.canvas.
Materializes literature sources into local references/ directory using Paperpile for academic papers, Readwise for personal highlights, and Obsidian for web clips after brainstorm phase.
Share bugs, ideas, or general feedback.
Bridge between research-hub and academic-writing-skills. Reads a
manuscript draft (Word, LaTeX, Markdown, Obsidian paper folder) plus the
figures it references, and writes structured .paper/claims.yml +
.paper/figures.yml so the writing skill can reason about the paper
without parsing the manuscript on every call.
This skill does not edit the manuscript itself. That's the writing skill's job. We just produce the memory layer it consumes.
Trigger phrases:
Not for:
academic-writing-skills.research-project-orienter.literature-triage-matrix.notebooklm-brief-verifier.In priority order:
.paper/ existing files — if claims.yml already exists, parse
it; this run should refresh, not replace human edits.paper/manuscript.docx / manuscript.tex / manuscript.mdraw/<paper-cluster>/manuscript.mdfigures/ or outputs/figures/ — read
filenames + captions; don't OCR images..research/project_manifest.yml — for project context (research
question, datasets, current_stage). Use to anchor claims in the
project's overall question..research/literature_matrix.md — for citation key lookup if the
manuscript references "Smith 2024" but you need to disambiguate.Write to <project-root>/.paper/:
claims.yml — every paper-level claim, with evidence pointers + status.figures.yml — every figure inventory, with key numbers + supported
claims.revision_history.yml — append-only log of revision rounds (when a
claim or figure changed, in which round, why). See
references/revision_history_schema.md for the schema and the
append-vs-overwrite rules.Do not touch journal_format.md, reviewer_comments.md, or
style_overrides.md — those belong to academic-writing-skills.
.paper/claims.yml structureclaims:
- id: C1
text: "Coupled ABM-CAT reduces flood-impact RMSE by 22%."
evidence_artifacts:
- "outputs/E2/calibration.csv"
- "outputs/E2/figure3.png"
figure_or_table: ["Fig3"]
status: "draft" # draft | supported | rejected
risk: "Reviewer R2 may push back on calibration window."
sentence_in_manuscript: "...we observe a 22% reduction in RMSE..."
Required per claim: id, text, status. Numbering: C1, C2, ...
contiguous; if you regenerate, preserve human-assigned IDs.
.paper/figures.yml structurefigures:
- id: "Fig1"
file: "outputs/figures/Fig1_study_area.png"
panels: ["a) site map", "b) gauge locations"]
key_numbers: ["12 gauges", "1985-2024 record length"]
supports_claims: []
caption_in_manuscript: "Figure 1. Study area..."
Required per figure: id, file, supports_claims (may be []).
.paper/revision_history.yml structurerevisions:
- round: 1
date: "2026-04-15"
trigger: "Initial draft v1"
changed_claims: [] # all claims new
changed_figures: []
summary: "First complete draft."
- round: 2
date: "2026-05-02"
trigger: "Reviewer 2 round 1: doubt on calibration window"
changed_claims: ["C1", "C4"]
changed_figures: ["Fig3"]
summary: "Tightened calibration window claim, added robustness check (Fig S3), softened C4 claim wording."
Required per revision: round, date, trigger, summary.
changed_claims and changed_figures may be [] for fresh drafts.
Append, do not overwrite. Each new revision round appends a new
list entry. Past rounds stay verbatim — they're the audit trail. The
full schema is in references/revision_history_schema.md.
.paper/*.yml in subsequent writing turns.academic-writing-skills next for any
writing/revision/audit pass."[paper-memory-builder]
Read manuscript: paper/manuscript.docx (8 pages, 12 figures)
Wrote: .paper/claims.yml (9 claims; 2 marked at risk)
Wrote: .paper/figures.yml (12 figures; 8 mapped to claims, 4 are context)
Suggested next: load academic-writing-skills for revision/audit passes.
.paper/journal_format.md, reviewer_comments.md, or
style_overrides.md. Those are owned by academic-writing-skills..research/ — that's the workspace layer, not the
paper layer.