agentleFS
flonatflonat/flonat-research/skills/insights-deck/SKILL.md

insights-deck

Skill150 starsChanged 10 days agoMITView on GitHub

Archive an exported Claude Code insights HTML report and turn it into a timestamped Beamer presentation usable from either client. Use when the insights export is the source artifact for a presentation. Not for a general research talk; use $talk-deck.

The file

We split this skill into the rules it gives an agent

9 rules. About 0 are common, 6 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    After exporting the built-in Claude Code insights report

    line 14
  2. Distinctive:

    When you want a shareable deck summarising how you use Claude Code

    line 15
  3. Distinctive:

    If an HTML path was supplied, continue directly.

    line 32
  4. Distinctive:

    After the export completes, find the most recent insights HTML file:

    line 36
  5. Very specific:

    If not found in /tmp/, check the user's home directory and common download locations.

    line 42
  6. Distinctive:

    Use today's date for the timestamp.

    line 53
  7. Very specific:

    Read the archived HTML file (log/insights/YYYY-MM-DD/insights-YYYY-MM-DD-log.html).

    line 61
  8. Very specific:

    Write to log/insights/YYYY-MM-DD/insights-YYYY-MM-DD-deck.tex with:

    line 83
  9. Distinctive:

    When complete, the date folder should contain:

    line 128

≈ “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

Insights Deck Skill

Archive a Claude Code insights HTML export and generate a Beamer presentation summarising the findings. All outputs go to log/insights/ with date stamps.

When to Use

  • After exporting the built-in Claude Code insights report
  • When you want a shareable deck summarising how you use Claude Code
  • For periodic self-reflection on AI-assisted workflow

Scope

This skill produces two outputs only: an archived HTML report and a Beamer deck. Do NOT use insights suggestions to create plans, rules, CLAUDE.md edits, or other follow-up actions during this skill's execution. If the user wants to act on suggestions, that happens after the deck is delivered — as a separate conversation.


Phase 1 — Generate & Archive Insights

Step 1: Obtain the exported HTML

The insights export is produced by a built-in Claude Code command and cannot be generated by Codex. If no HTML path was provided, tell the user:

Please export the Claude Code insights report, then provide the generated HTML path. I can build the archive and deck from either client.

If an HTML path was supplied, continue directly. Do not require the active client itself to be Claude Code.

Step 2: Locate the generated HTML

After the export completes, find the most recent insights HTML file:

ls -t /tmp/claude-insights-*.html 2>/dev/null | head -1

If not found in /tmp/, check the user's home directory and common download locations. Ask the user for the path if it can't be located automatically.

Step 3: Archive the HTML

Each insights run gets its own date folder:

mkdir -p log/insights/YYYY-MM-DD
cp <source-html> log/insights/YYYY-MM-DD/insights-YYYY-MM-DD-log.html

Use today's date for the timestamp.


Phase 2 — Build Beamer Deck

Step 1: Read and extract findings

Read the archived HTML file (log/insights/YYYY-MM-DD/insights-YYYY-MM-DD-log.html). Extract:

  1. Usage patterns — most-used tools, session frequency, typical session length
  2. Strengths — what's working well in the workflow
  3. Friction points — repeated failures, slow patterns, underused features
  4. Recommendations — suggested improvements

Step 2: Design the deck structure

Apply rhetoric principles from the beamer-deck skill (condensed — no sub-agent reviews needed for an internal deck):

  • Assertion titles — every frame title states a claim, not a label
  • One idea per slide — split if a slide has two ideas
  • Three-act arc:
    • Act I (Tension): Current usage snapshot — what does the data show?
    • Act II (Development): Patterns, strengths, friction points
    • Act III (Resolution): Recommendations and next steps
  • MB/MC balance — vary dense and light slides deliberately
  • Pyramid principle — lead with conclusions, then support

Step 3: Generate the .tex file

Write to log/insights/YYYY-MM-DD/insights-YYYY-MM-DD-deck.tex with:

  • \documentclass[aspectratio=169,11pt]{beamer}
  • Original inline theme (no default Beamer themes)
  • Professional colour palette:
    \definecolor{Midnight}{HTML}{1A1A2E}
    \definecolor{DeepBlue}{HTML}{16213E}
    \definecolor{RoyalBlue}{HTML}{0F3460}
    \definecolor{Coral}{HTML}{E94560}
    \definecolor{SoftGray}{HTML}{BDC3C7}
    \definecolor{CloudWhite}{HTML}{FAFBFC}
  • Suggested structure (adapt based on findings):
    1. Title slide — "Claude Code Usage: [Month Year]"
    2. Opening claim — lead with the most striking finding
    3. Usage overview — session count, tool distribution, time patterns
    4. Top strengths (1–2 slides)
    5. Key friction points (1–2 slides)
    6. Recommendations (1–2 slides)
    7. Closing — single actionable takeaway

Step 4: Compile

Install the fail-closed canonical .latexmkrc in the date folder if it is missing by invoking the latex pre-flight, which resolves its bundled canonical artifact. Never reconstruct a shortened recipe. If a file already exists and differs, classify it before migration.

Then compile:

cd log/insights/YYYY-MM-DD && latexmk -pdf insights-YYYY-MM-DD-deck.tex

Step 5: Fix all warnings

Parse out/*.log for overfull/underfull hbox/vbox warnings. Fix every one. Recompile until clean.

The canonical .latexmkrc copies the requested PDF from out/ back to the date folder only after a successful build.


Output

When complete, the date folder should contain:

log/insights/YYYY-MM-DD/
├── insights-YYYY-MM-DD-log.html     # Archived insights HTML
├── insights-YYYY-MM-DD-deck.tex     # Beamer source
├── insights-YYYY-MM-DD-deck.pdf     # Compiled PDF
├── .latexmkrc                       # Canonical build config
└── out/                             # Build artifacts

Each insights-deck run creates a new date folder, keeping the parent log/insights/ clean.


Cross-References

SkillRelationship
beamer-deckFull rhetoric framework and multi-agent review (this skill uses a condensed version)
latexDefault compiler — use for compilation with auto error resolution
latexManual compilation config details, .latexmkrc setup
session-logComplements insights with session-level detail

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.