agentleFS
Jawad-OfficialJawad-Official/writing-toolkit-skills/skills/work-report-writer/SKILL.md

work-report-writer

Skill1 starsChanged 35 days agoMITView on GitHub

Use for reports at work from his data (analysis reports, status updates, proposals, meeting summaries, memos) for a manager, client, or team. Asks for a text style, then writes full drafts, executive summary first, numbers checked.

The file

Show its rules and text

We split this skill into the rules it gives an agent

13 rules. About 0 are common, 12 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈12

    Distinctive · estimate

    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:

    Write workplace reports from the user's data.

    line 8
  2. Distinctive:

    If the writing-core skill is available, load it first.

    line 12
  3. Very specific:

    Never invent data, results, names, or dates.

    line 18
  4. Distinctive:

    Write in the text style he picked (Step 0.5).

    line 19
  5. Distinctive:

    No style changes the rules: nothing invented, markers for gaps, no em dashes, no moral endings.

    line 47
  6. Distinctive:

    Ask only for what is missing, in one message, at most 3 questions:

    line 51
  7. Distinctive:

    Compute, do not estimate: When he gives files (CSV, XLSX, exports), calculate totals, rates, and changes with code, not by reading the …

    line 63
  8. Distinctive:

    Use the variant that fits the request.

    line 83
  9. Distinctive:

    If his company has a template, follow it over these defaults.

    line 91
  10. Distinctive:

    If he asks you to build charts, load the dataviz skill first.

    line 96
  11. Distinctive:

    Use a table when the reader needs exact values.

    line 97
  12. Distinctive:

    Long reports: if the report runs over about 1,000 words, or he asks for a file, build a .docx with the docx skill.

    line 102
  13. Distinctive:

    No edit guide and no teaching notes by default.

    line 104

≈ “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

Work Report Writer

Write workplace reports from the user's data. He provides the data and context, Claude writes the full draft, and he checks it and sends it. Unlike university work, the reader cares about accuracy, clarity, and the decision, not about who typed the sentences.

Step 0: Load the core rules

If the writing-core skill is available, load it first. It holds the full shared rules (his inputs first, avoiding generic AI defaults, never inventing, markers, voice, self-check). If it is not available, follow this summary:

  • State each point once, early. No closing sentences that restate.
  • Name specifics: numbers, dates, tools, sources, people's roles.
  • Include real limits, risks, and trade-offs from his material.
  • Plain literal language. No decoration, no em dashes, no filler words like delve, landscape, pivotal, crucial.
  • Never invent data, results, names, or dates. Use markers: [YOUR INPUT: ...], [VERIFY: ...], [FIGURE: ...], [CITE: ...].
  • Write in the text style he picked (Step 0.5).
  • Take a position. End on the recommendation or next step, not a moral.

Step 0.5: Ask for the text style

At the start of every new task, ask which text style he wants. Show all 12 styles as a numbered list in the same message as the intake questions, and mark the recommended ones. He can pick one number or combine two (for example "2 + 4"). The style question does not count toward the 3-question intake limit.

Recommended for this skill: 3 Professional for most reports and emails. 7 Concise executive for status updates, memos, and summaries. 6 Persuasive for proposals. 4 Technical for engineering readers.

Skip the question when he already named a style or tone, when he says "same style" or "you choose", or when the session runs unattended. Then use the closest style, or the first recommended one, and say which in one short line. Follow-up edits on the same piece keep the chosen style. If he skips the question, use the first recommended style and do not ask again for that task.

The full definitions are in writing-core (Rule 0). If it is not loaded, use this list exactly:

1. Plain and clear: common words, short to medium sentences. The default.
2. Formal academic: impersonal, no contractions, precise hedging.
3. Professional: direct and polite, answer first.
4. Technical: exact terms, versions, units, and names.
5. Conversational: friendly, contractions, speaks to "you".
6. Persuasive: claim, evidence, objection answered, clear ask.
7. Concise executive: shortest form, numbers and decisions first.
8. Reflective personal: first person, named feelings, what he learned.
9. Teaching: step by step, terms defined, example before the rule.
10. Storytelling: hook, his real story, tension, payoff.
11. Simple English: very common words, no idioms, for non-native readers.
12. Match my writing: mirrors a sample he pastes.

No style changes the rules: nothing invented, markers for gaps, no em dashes, no moral endings. Storytelling and Reflective use only his real stories and feelings.

Step 1: Intake

Ask only for what is missing, in one message, at most 3 questions:

  1. Reader and purpose. Who reads it (manager, client, team, board) and what they need to do after reading (approve, decide, stay informed, fix something).
  2. Data and context. Files, numbers, notes, previous reports, the time period covered, and any company template.
  3. His position. What he thinks the recommendation should be, if he has one.

Also note length limits and the deadline if given.

Step 2: Handle the data carefully

Numbers are where a work report can do real damage, so this step comes before any writing.

  • Compute, do not estimate. When he gives files (CSV, XLSX, exports), calculate totals, rates, and changes with code, not by reading the table by eye. Show the calculation in the reply or the appendix.
  • State the source of every number: which file, sheet, or system, and the date range.
  • Flag problems, do not smooth them over: missing values, outliers, a sudden jump, mismatched totals, small sample sizes. Put them in the report as limits, or ask him first.
  • Units and comparisons: say what each number is compared to (last month, target, same period last year). Distinguish percent change from percentage points.
  • Personal data: if the data contains names or details of students, customers, or employees, remind him once to anonymize before sharing, unless he already did. Do not repeat personal details in the report unless the report needs them. Also remind him once to check his workplace's AI policy if he has not mentioned it.

Step 3: Structure (answer first)

Readers at work skim. Put the answer first, then the support (the pyramid principle). Default structure for an analysis report:

  1. Title that says the finding, not only the topic. For example "Lesson completion fell 12% after the March update", not "Completion report".
  2. Executive summary (3 to 5 sentences): the situation, the key finding with its number, the recommendation, and the decision or action needed from the reader.
  3. Findings. Each finding starts with a headline sentence that contains the number, then the evidence, then why it matters. Most important finding first.
  4. Recommendation. The options considered, the trade-off of each, and the recommended option with the reason. If he has not decided, present the options, mark the one the data supports, and add [YOUR INPUT: confirm your recommendation]. The decision is his.
  5. Risks and limits. Data gaps, assumptions, what could make the recommendation wrong.
  6. Next steps. Each with an action, an owner, and a date. Use [YOUR INPUT: owner] and [YOUR INPUT: date] where unknown.
  7. Appendix. Method, data sources, calculations, full tables.

Variants

Use the variant that fits the request. Keep the answer-first rule in all of them.

  • Status or progress update: done since the last update, in progress, blocked (with what would unblock it), next, and asks for the reader. Lead with overall status in one line (on track, at risk, or off track) and the reason.
  • Proposal: the problem with evidence, the proposed solution, cost and time, alternatives considered, risks, and the specific ask (budget, approval, people).
  • Meeting summary: decisions made, action items (action, owner, date), open questions, and a short context line. Skip the play-by-play of who said what.
  • One-page memo: the bottom line in the first sentence, three supporting points with numbers, and the ask.
  • Email report: subject line with the key finding, the summary in the first two lines, details below, and attachments named.

If his company has a template, follow it over these defaults.

Charts and tables

  • Suggest one chart per key finding at most, with [FIGURE: ...] saying the chart type and what it shows. For example "line chart of weekly completion rate, Jan to Jun, with the March update marked".
  • If he asks you to build charts, load the dataviz skill first.
  • Use a table when the reader needs exact values. Use a chart when the reader needs the trend or comparison.

Output format

  • Paste blocks: fenced text code blocks, one per section, labeled above the fence. No markdown inside. Note formatting to apply after each block (heading levels, bold, bullets).
  • Long reports: if the report runs over about 1,000 words, or he asks for a file, build a .docx with the docx skill. Build an .xlsx with the xlsx skill if he wants the tables as a spreadsheet.
  • Check before sending: end every report with a short list for him: each number used, where it came from (file and row, column, or calculation), and any [VERIFY] items. This lets him check the numbers in two minutes before sending.
  • No edit guide and no teaching notes by default. He can ask for them.

Never do these

  • Invent or round numbers in a way that changes meaning.
  • Hide bad results or data problems.
  • Make the decision for him without saying it is a draft recommendation.
  • Include personal data the report does not need.
  • Pad with background the reader already knows.

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.