Vis Lens Always On

Always-On Visualization Triage Lens

Philosophical Mode: Composite Primary Question: "What are the blocking visualization issues?" Focus: Anti-Pattern Detection, Accessibility, Annotation Completeness — combined triage

Arguments

/autoskillit:vis-lens-always-on [context_path] [experiment_plan_path]

• context_path (optional positional arg 1) — Absolute path to a lens context file containing IV/DV tables, H0/H1 hypotheses, controlled variables, and success criteria. If provided, read this file before beginning analysis to obtain structured context. If omitted, discover context by exploring the CWD.
• experiment_plan_path (optional positional arg 2) — Absolute path to the full experiment plan. If provided, read for complete experimental methodology and design. If omitted, locate the experiment plan by exploring the CWD.

When to Use

• Fast triage before any figure review or submission — run this first
• Lightweight check when you do not need the full depth of individual lenses
• CI-style gate: FAIL_N blocks, WARN_N is informational, PASS is green
• Quick scan of a figure plan to surface the most important blocking issues
• User invokes /autoskillit:vis-lens-always-on

Critical Constraints

NEVER:

• Modify any source code files
• Do not litter the codebase with useless comments, TODO markers, or explanatory annotations — the skill output and diagram speak for themselves
• Create files outside {{AUTOSKILLIT_TEMP}}/vis-lens-always-on/
• Emit PASS if any critical finding exists — any single critical = FAIL_N

ALWAYS:

• Run all three passes in sequence; do not skip any pass

• Tally critical_count and warning_count across all three passes

• BEFORE creating any diagram, LOAD the /autoskillit:mermaid skill using the Skill tool - this is MANDATORY

• If the Skill tool cannot be used (disable-model-invocation) or refuses this invocation, do NOT proceed with diagram creation. Abort this step and omit the diagram from output.

• Write output to {{AUTOSKILLIT_TEMP}}/vis-lens-always-on/vis_spec_always_on_{YYYY-MM-DD_HHMMSS}.md (relative to the current working directory)

• After writing the file, emit the structured output token as literal plain text with no markdown formatting on the token name (the adjudicator performs a regex match):

  diagram_path = /absolute/path/to/{{AUTOSKILLIT_TEMP}}/vis-lens-always-on/vis_spec_always_on_{...}.md
  

Analysis Workflow

Step 0: Parse optional arguments

If positional arg 1 (context_path) is provided and the file exists, read it to obtain IV/DV tables, H0/H1 hypotheses, controlled variables, and success criteria. If positional arg 2 (experiment_plan_path) is provided and exists, read the experiment plan for full methodology. Use this structured context as the foundation for all three passes; skip the CWD exploration for these fields if the context file supplies them.

Step 1: Inventory All Figures

Before running passes, build a complete figure inventory from the experiment plan, context file, code, and any existing output files. Every figure in this inventory is checked in all three passes.

Step 2: Run THREE Sequential Passes

Pass 1 — Anti-Pattern Scan

Check all figures against the 16 anti-patterns from vis-lens-antipattern:

Tally: pass1_critical_count, pass1_warning_count

Pass 2 — Accessibility Scan

For each figure, check all four accessibility criteria:

1. colorblind_safe — palette is wong, okabe-ito, viridis, or cividis? (PASS) or jet/rainbow/custom? (FAIL → warning)
2. font_size_ok — all axis labels, tick labels, legends ≥ 8pt? Check fontsize args in code or figure config. (PASS if ≥ 8pt or undetectable; FAIL → warning)
3. no_color_only_encoding — distinction between groups uses shape, pattern, or label in addition to color? Color-only encoding for distinction = FAIL → warning
4. captions_present — figure has alt-text, caption, or description in the paper/plan? No caption = FAIL → warning

Tally: count failures as pass2_warning_count (accessibility failures are warnings, not critical)

Pass 3 — Annotation Completeness

For each figure, check all five annotation criteria:

1. all_titles — every figure has a title or caption title? Missing = warning
2. axis_labels — both axes labeled with variable name and units? Missing = warning
3. legends_ok — multi-series figures have a legend? Missing = warning
4. stat_overlay_labeled — if an uncertainty overlay exists, is it labeled (e.g., "± CI95, n=5")? Unlabeled overlay = warning
5. data_pointer — figure has a data source reference or code pointer in caption? Missing = info (not counted in tally)

Tally: count failures as pass3_warning_count

Step 3: Compute Combined Verdict

critical_count = pass1_critical_count
warning_count = pass1_warning_count + pass2_warning_count + pass3_warning_count

if critical_count > 0:
    verdict = f"FAIL_{critical_count}"
elif warning_count > 0:
    verdict = f"WARN_{warning_count}"
else:
    verdict = "PASS"

Step 4: Emit yaml:spec-index Block and Mermaid Diagram

Emit the yaml:spec-index triage index (NOT yaml:figure-spec — this is a triage summary, not a per-figure spec). Then LOAD /autoskillit:mermaid and create the triage diagram.

Output Template

# Always-On Triage Report: {System / Experiment Name}

**Lens:** Always-On Triage (Composite)
**Question:** What are the blocking visualization issues?
**Date:** {YYYY-MM-DD}
**Scope:** {What was analyzed}

## Triage Index

```yaml
# yaml:spec-index — Always-On triage index
verdict: "FAIL_3"         # PASS | WARN_N | FAIL_N
critical_count: 3
warning_count: 5
pass_1_antipattern: ["ap-single-seed", "ap-rainbow", "ap-dual-axis"]
pass_2_accessibility:
  colorblind_safe: true
  font_size_ok: false
  no_color_only_encoding: true
  captions_present: false
pass_3_annotation:
  all_titles: true
  axis_labels: false
  legends_ok: true
  stat_overlay_labeled: false
  data_pointer: true
figures_with_critical: ["fig-01-main-results", "fig-03-ablation"]

Pass Results

Pass 1 — Anti-Pattern Scan

Pass 2 — Accessibility Scan

Pass 3 — Annotation Completeness

Triage Diagram

%%{init: {'flowchart': {'nodeSpacing': 50, 'rankSpacing': 60, 'curve': 'basis'}}}%%
flowchart TB
    %% CLASS DEFINITIONS %%
    classDef cli fill:#1a237e,stroke:#7986cb,stroke-width:2px,color:#fff;
    classDef stateNode fill:#004d40,stroke:#4db6ac,stroke-width:2px,color:#fff;
    classDef handler fill:#e65100,stroke:#ffb74d,stroke-width:2px,color:#fff;
    classDef newComponent fill:#2e7d32,stroke:#81c784,stroke-width:2px,color:#fff;
    classDef output fill:#00695c,stroke:#4db6ac,stroke-width:2px,color:#fff;
    classDef detector fill:#b71c1c,stroke:#ef5350,stroke-width:2px,color:#fff;
    classDef gap fill:#ff6f00,stroke:#ffa726,stroke-width:2px,color:#000;

    subgraph Pass1 ["PASS 1: ANTI-PATTERN"]
        P1["{N} critical<br/>━━━━━━━━━━<br/>{ap-ids found}"]
    end

    subgraph Pass2 ["PASS 2: ACCESSIBILITY"]
        P2["{N} warnings<br/>━━━━━━━━━━<br/>{failed checks}"]
    end

    subgraph Pass3 ["PASS 3: ANNOTATION"]
        P3["{N} warnings<br/>━━━━━━━━━━<br/>{failed checks}"]
    end

    VERDICT["{FAIL_3 / WARN_5 / PASS}<br/>━━━━━━━━━━<br/>critical={N}, warning={M}"]

    P1 --> VERDICT
    P2 --> VERDICT
    P3 --> VERDICT

    class P1 detector;
    class P2 gap;
    class P3 handler;
    class VERDICT cli;

Color Legend:

---

## Pre-Diagram Checklist

Before creating the diagram, verify:

- [ ] LOADED `/autoskillit:mermaid` skill using the Skill tool
- [ ] Using ONLY classDef styles from the mermaid skill (no invented colors)
- [ ] Diagram will include a color legend table
- [ ] All three passes completed before computing verdict
- [ ] yaml:spec-index emitted (not yaml:figure-spec)
- [ ] verdict is FAIL_N if critical_count > 0, WARN_N if warning_count > 0, PASS otherwise