agentleFS
luisgui1757luisgui1757/sentinel/skills/tldr/SKILL.md

tldr

Skill0 starsChanged 10 days agoMITView on GitHub

Load this skill before you shorten a reply, a document, a file or a thread that exists. Use it when the user asks for a tldr, a TLDR, a summary or a recap. Use it also for the key points or the short version of a text. Use it also when the user says that a text is too long. Use it also when the user has no time to read a text. Do not use it to explain a command, a concept or code. Do not use it to write the summary section of a document or a PR. The skill keeps each decision, option, necessary change, doubt and limit of the source.

40 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use explain, handoff and 2 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • sunnyclinches44Skill

    explain

    Explains code using real-world analogies and ASCII diagrams

    Also in 3 projects that use this

  • MicrosoftSkill

    handoff

    Compact the current conversation into a handoff document for another agent to pick up.

    Also in 6 projects that use this

  • Google CloudSkill

    triage

    Gather context on a change or bug and plan the implementation.

    Also in 3 projects that use this

  • GoogleSkill

    grill-me

    A relentless interview to sharpen a plan or design.

    Also in 3 projects that use this

The file

Show its rules and text

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 35 edited versions we read keep these.

  • 14

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use the text that the user names.

    line 14
  2. Distinctive:

    Start with the answer. Then give one line for each item, in the order of importance, with each part that the source gives.

    line 38
  3. Distinctive:

    Write an option and a necessary change in these shapes.

    line 44
  4. Distinctive:

    When the source has several decisions or options, put the lines in groups.

    line 52
  5. Distinctive:

    Keep the verb of obligation: "must", "should", "suggests", "can".

    line 63
  6. Distinctive:

    Keep who says, doubts or does not know each thing, and whom it covers.

    line 65
  7. Distinctive:

    Keep the scope of each claim: "only on X", "until Y", "if Z".

    line 68
  8. Distinctive:

    Keep each part of a change.

    line 69
  9. Distinctive:

    Keep numbers, names, identifiers, code and quotations exactly.

    line 71
  10. Distinctive:

    Do not add a note, a cause, a conclusion or a computed number of your own.

    line 75
  11. Distinctive:

    Do not fill a part that the source leaves open.

    line 76
  12. Distinctive:

    Remove narration, background, praise, repeated text and offers of more work.

    line 81
  13. Distinctive:

    Remove the method of a test, and how a step or a tool works inside.

    line 82
  14. Distinctive:

    Remove a check that did not run when its result changes nothing that the reader does.

    line 84

“Common” is measured: kept by at least 6 in 10 of the 35 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

TLDR

Shorten a text that exists. Keep what the reader needs for a decision or an action. The number of these items sets the length. The length of the source and a word count do not.

Target

Use the text that the user names. If the user names no text, use your last reply. Read the full source before you write.

List the items

List each item of the source that changes what the reader decides or does. For each item, look for these parts in the source:

  • The answer or the recommendation, with its strength and its main reason.
  • Each decision that waits on the reader, and who decides. For each option, the exact change, what it gives, what it costs or loses, and what it needs.
  • Each change that must happen, with each of its parts and values, the date, and the result when it does not happen.
  • Each condition or limit of a claim, a plan or an action. Look also in the sentences that name no problem.
  • Each doubt or open question, with who holds it, its evidence, and who does not know the answer.
  • Each check that did not run, or that ran only in part.
  • Each number, name, date or cost that the reader needs.

This list is for you. A part that the source does not give stays out.

Write the result

Start with the answer. Then give one line for each item, in the order of importance, with each part that the source gives. Put a reason or another needed fact in the line of the item that it supports. Each sentence gives an item or a part of your list. Remove a sentence that gives none, and do not remove an item or a part to save words.

Write an option and a necessary change in these shapes. Leave out a part that the source does not give.

- <option>: <the exact change>. Gives: <what>. Costs: <what>. Needs: <what>.
- <change>: <each part, with its values>. By <date>. If it does not happen: <result>.

When the source has several decisions or options, put the lines in groups. Give each group a label, such as "Decision", "Must do", "Limits" or "Open". Give each option its own line. For a yes or no decision, give what each answer causes, when the source says it.

Give each fact once. Write whole sentences.

Keep the meaning

  • The certainty of the source is a fact, not a hedge. Keep its words, such as "seems", "probably", "slightly" or "not sure".
  • Keep the verb of obligation: "must", "should", "suggests", "can". Keep a plain statement plain. "The test fails" is not "the test can fail".
  • Keep who says, doubts or does not know each thing, and whom it covers. "I do not know" is not "nobody knows". "One reviewer disagrees" is not "the reviewers disagree".
  • Keep the scope of each claim: "only on X", "until Y", "if Z".
  • Keep each part of a change. "Set the timeout to 30 s" is a change. "Fix the timeout" is only a name for one.
  • Keep numbers, names, identifiers, code and quotations exactly.

Add nothing

Do not add a note, a cause, a conclusion or a computed number of your own. Do not fill a part that the source leaves open. Do not write about your own work, such as what you read or did not change.

Remove the rest

Remove narration, background, praise, repeated text and offers of more work. Remove the method of a test, and how a step or a tool works inside. Remove a detail that only explains an item, such as the line count behind a slow test. Remove a check that did not run when its result changes nothing that the reader does.

Check

Compare the source with the result, paragraph by paragraph. Find each item of your list and each of its parts. Then compare each sentence of the result with the source. It must have the same strength, scope and owner. Restore what is missing or changed, in the words of the source.

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.