agentleFS
JipletJiplet/transformation-os/skills/program/program-exec-summary/SKILL.md

exec-summary

Skill2 starsChanged 4 months agoMITView on GitHub

Compress any analysis into a structured one-page executive summary as a Word document, formatted for CFO, ExCo, or senior stakeholder audiences.

7 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

7 rules. 0 are common, 6 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 6

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    After completing analysis that needs to be handed off to a senior audience

    line 17
  2. Distinctive:

    Every Evidence bullet has a number: If you can't attach a figure, cut the bullet.

    line 88
  3. Distinctive:

    One page. Hard limit: If it doesn't fit, cut content — never shrink font or margins.

    line 90
  4. Distinctive:

    Never fabricate data or numbers — if source material is thin, say so in the Finding rather than manufacturing one

    line 98
  5. Distinctive:

    If the source analysis doesn't support a clear position, write: "Finding: Insufficient data to draw a conclusion — [specific gap]" and …

    line 99
  6. Distinctive:

    Always mark DRAFT unless Jacob explicitly says FINAL

    line 100
  7. Very specific:

    After delivering the output, compile durable findings to the Knowledge Wiki.

    line 119

“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “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

Skill: exec-summary

Purpose

Takes any analysis output — from conversation, a file, or another skill's output — and compresses it into a structured one-page executive summary. Output is always a .docx file using python-docx.

When to Invoke

Trigger phrases (exact or close variants):

  • "exec summary", "one-pager", "summarise this for [audience]"
  • "CFO summary", "ExCo brief", "make this executive-ready"
  • After completing analysis that needs to be handed off to a senior audience

One-Pager Structure

TITLE — [Topic]
Date | Status: DRAFT / FINAL | Prepared by: Transformation

1. SITUATION (2–3 sentences)
   Context: what is this, and why are we looking at it now?

2. FINDING (1–2 sentences — bolded)
   The single most important takeaway. Takes a position.
   No hedging. "X costs $Y more than benchmark" not "data suggests X may be higher."

3. EVIDENCE (3–5 bullets)
   Key facts supporting the finding. Every bullet must contain a number.
   No qualitative-only points.

4. IMPLICATIONS (2–3 bullets)
   What this means for the business.
   Connect to 8 figure FY26 / 9 figure 3-year target where relevant.

5. RECOMMENDATION (numbered list)
   Specific, assignable actions with owner and timeframe.
   Each action labelled: Quick Win / Requires Approval / Structural
   "Investigate X" is not a recommendation. "Jacob to present X to CFO by [date]" is.

6. RISKS & DEPENDENCIES (table)
   Columns: Risk | Impact | Mitigation
   Max 3–4 rows. Material risks only.

Document Generation

  • Library: python-docx only — never docx-js (won't open in Word for Mac)
  • Venv: <workspace>/.venv
  • Run pattern: source <workspace>/.venv/bin/activate && python3 - << 'EOF' ... EOF
  • Save path: 4 - Projects/[project-name]/[YYMMDD]_Exec_Summary_[topic].docx
  • Logo: Include the configured brand logo top-left if provided (bring your own theme)

Brand Specifications

Brand-adaptable output: headings, section dividers, and body font follow the active configured theme rather than a hardcoded palette.

ElementValue
CurrencyAUD, $#,##0 format (no decimals)
Status defaultDRAFT (unless Jacob says FINAL)

Execution Steps

  1. Identify source material — use the most recent analysis in conversation, or load the file Jacob specifies. If ambiguous, ask one scoping question: "Which output should I summarise, and who is the audience?"
  2. Draft structure in conversation first — show the six sections as plain text before generating the docx. Wait for Jacob to confirm or adjust before writing the file.
  3. Generate the docx — use inline python-docx heredoc via Bash. Apply the configured brand colours to headings, bold the Finding, render Risks as a Word table with a thin border in the brand accent colour.
  4. Save and confirm — report the save path and flag any data gaps or unverified numbers.

Design Principles

  • Lead with the Finding. It is the most important section. It takes a clear position — not a hypothesis, not a hedge.
  • Every Evidence bullet has a number. If you can't attach a figure, cut the bullet.
  • Recommendations are assignable. Name an owner and a date. If you don't know the owner, flag it — don't leave it vague.
  • One page. Hard limit. If it doesn't fit, cut content — never shrink font or margins.
  • 60-second readability. A time-poor CFO should finish it standing at a printer.
  • No AI language. No "this analysis leverages", no "it is worth noting". Write like a project director briefing ExCo over coffee.

Rules

  • Never fabricate data or numbers — if source material is thin, say so in the Finding rather than manufacturing one
  • If the source analysis doesn't support a clear position, write: "Finding: Insufficient data to draw a conclusion — [specific gap]" and recommend the data action
  • Always mark DRAFT unless Jacob explicitly says FINAL
  • Connect Implications and Recommendations to the multi-year cost program where there is a genuine link — do not force it
  • Append a dated entry to 4 - Projects/[project-name]/log.md on completion

Programme Context

  • Company: a large ASX-listed infrastructure services enterprise, Australian infrastructure services
  • Sectors: Defence & Social Infrastructure (D&SI), Telco, Transport, Infrastructure Services
  • Cost program: 8 figure FY26 productivity target / 9 figure 3-year transformation
  • Default audience: CFO, ExCo, or sector leadership, senior, time-poor, decision-ready
  • Competitive set: major Australian infrastructure services peers
  • All figures AUD unless stated

Wiki Compile (post-delivery)

After delivering the output, compile durable findings to the Knowledge Wiki. Read context/wiki-compile-step.md for the full checklist. Skip if the output is formatting-only or contains no new findings (apply the "so what" test).

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.

Reviews can't be read right now.

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.