agentleFS
169884902hzl169884902hzl/engineering-paper-skills/skills/engineering-figure-table/SKILL.md

engineering-figure-table

Skill3 starsChanged 40 days agoMITView on GitHub

Plan, revise, audit, or polish figures, tables, captions, table notes, and visual evidence flow for English engineering conference or journal papers. Use when the user asks about figure responsibility, multi-panel layout, main-result tables, ablation tables, captions, table formatting, object/category naming consistency, page budget for visuals, or whether visuals support manuscript claims.

The file

Show its rules and text

We split this skill into the rules it gives an agent

14 rules. About 0 are common, 10 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈10

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill to make figures and tables serve the manuscript argument.

    line 8
  2. Distinctive:

    Do not let captions claim what the visual does not show.

    line 15
  3. Distinctive:

    Do not use tables as unstructured parameter dumps unless the table's job is setup compression.

    line 18
  4. Distinctive:

    Keep category names, metric names, captions, and prose consistent.

    line 20
  5. Very specific:

    Use engineering-writing when the user needs section prose around the visual evidence.

    line 35
  6. Very specific:

    Use engineering-polishing for prose-only clarity or style work.

    line 37
  7. Very specific:

    Use engineering-response when visual changes are driven by reviewer or advisor comments.

    line 38
  8. Very specific:

    Use engineering-validation for final label/reference/build consistency before claiming submission readiness.

    line 40
  9. Distinctive:

    Check source-data and caption consistency when values, metrics, or panels are provided.

    line 68
  10. Distinctive:

    Check whether the caption states only visible or tabulated information.

    line 70
  11. Distinctive:

    Check consistency across prose, caption, labels, and notes.

    line 71
  12. Distinctive:

    Check whether any category/metric rename needs coordinated changes.

    line 72
  13. Distinctive:

    Remove or redesign a visual if it has no claim, setup fact, evidence axis, or reader action.

    line 73
  14. Distinctive:

    If editing LaTeX, preserve labels and references unless there is a direct reason to change them.

    line 75

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Engineering Figure Table

Use this skill to make figures and tables serve the manuscript argument.

Core Stance

  • Define visual responsibility before visual styling.
  • A figure or table should carry a specific claim, setup fact, evidence axis, or reader action.
  • Do not let captions claim what the visual does not show.
  • Visuals can illustrate mechanisms, but they cannot prove hidden mechanisms without direct visual, tabular, or experimental evidence.
  • Do not use tables as unstructured parameter dumps unless the table's job is setup compression.
  • Keep category names, metric names, captions, and prose consistent.
  • Give each float one job. Showing the same data as a trend and as exact values is fine; showing it twice for the same purpose is not. Split the main comparison and the ablation when combining them makes either harder to read (P9 in reviewer-attack-patterns.md).
  • Make colors semantic (for example success versus failure), and keep one table style across the paper.
  • A teaser figure stays light, the framework figure carries the method, and qualitative result figures show representative cases chosen by a stated criterion, including observed failures (step 5 in writing-process.md).
  • Number floats in the order they are first referenced in the text (P7).

Boundaries

  • Use engineering-writing when the user needs section prose around the visual evidence.
  • Use engineering-polishing for prose-only clarity or style work.
  • Use engineering-response when visual changes are driven by reviewer or advisor comments.
  • Use engineering-validation for final label/reference/build consistency before claiming submission readiness.

When to Open Extra Files

FileOpen when
references/visual-contract.mdPlanning figure/table responsibility or deciding which visuals are needed
references/figure-roles.mdDeciding what motivation, framework, workflow, gallery, or evidence figures should do
references/captions.mdWriting or auditing captions and table notes
references/tables.mdDesigning setup, main results, ablation, or stress-test tables
references/panel-claim-map.mdMapping every panel, row, or column to a manuscript claim
references/figure-source-data-consistency.mdChecking that plotted values, source data, captions, and prose agree
references/venue-figure-standards.mdTarget venue family changes figure resolution, source data, or caption expectations
references/consistency.mdChecking category names, metrics, labels, caption style, and prose alignment
references/page-budget.mdCutting visual space, caption length, or table detail without damaging evidence
references/examples.mdNeeding concrete figure/table role examples and caption behavior
references/failure-modes.mdHandling captions or tables that claim more than the visual supports
../_shared/evidence-boundary.mdVisual claims may exceed visible or tabulated evidence
../_shared/claim-strength.mdCaption or table note verbs may overstate what is visible
../_shared/output-mode.mdThe user asks for caption/table output only
../_shared/terminology-ledger.mdCategory, metric, object, or method names change across visuals and prose

Workflow

  1. State the paper claim the visual supports.
  2. Assign one responsibility to each figure/table.
  3. Map each panel, row, or column to a piece of evidence.
  4. Check source-data and caption consistency when values, metrics, or panels are provided.
  5. Check whether the caption states only visible or tabulated information.
  6. Check consistency across prose, caption, labels, and notes.
  7. Check whether any category/metric rename needs coordinated changes.
  8. Remove or redesign a visual if it has no claim, setup fact, evidence axis, or reader action.
  9. If editing LaTeX, preserve labels and references unless there is a direct reason to change them.

Default Output

Visual audit
| Item | Responsibility | Visible/tabulated evidence | Claim supported | Must not claim | Placement | Related prose action | Risk | Action |

Panel-claim map
| Panel/row/column | Claim served | Evidence shown | Source-data check | Caption boundary |

Caption/table revision
[English caption or table plan]

Consistency actions
- labels:
- category names:
- metric names:
- prose references:

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.