tldr
Skill1 starsChanged 35 days agoMITView on GitHub
Compress the assistant's previous response into a high-signal summary — a bottom-line verdict plus icon-keyed bullets (✅ good / ⏳ in flight / ⬜ outstanding / ⚠️ trap / ❌ failed / ❓ unverified / ℹ️ context), each optionally carrying a short bold subject label, so a line's status and topic are both readable at a glance. **Pure compression — it deliberately ends with NO forward actions**: no Do-next, no action plan, no recommendation. Reach for /tldr when you want to understand what was said and decide for yourself what happens next. Can be narrowed with a focus argument. Chain with /recommend if an opinionated pick is also wanted.
39 other projects use a version of this.
Adopted by
Most recently taken up by
- richkuo/rk-skillstheir own version, first seen Nov 14, 2026
- annatomka/skillstheir own version, first seen Nov 8, 2026
- zazencodes/skillstheir own version, first seen Nov 8, 2026
- MarkPhamm/claude-skillstheir own version, first seen Nov 8, 2026
- nirberko/skillstheir own version, first seen Nov 8, 2026
- LaansDole/my-hermes-skillstheir own version, first seen Nov 7, 2026
Used with
Projects with it also use grill-me.
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.
GoogleSkillgrill-me
A relentless interview to sharpen a plan or design.
Also in 3 projects that use this
Other versions39 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 39
- earlyaidopters/second-brain★ 193same name
- saksham10arora-dotcom/everything-claude-codex★ 96same name
- richkuo/rk-skills★ 50same name
- gsarig/skills★ 30same name
- peterpme/skills★ 28same name
- Vesely/skills★ 28same name
- nirberko/skills★ 19same name
- r13i/skills★ 17same name
- zazencodes/skills★ 15same name
- SurefireStudios/tldr★ 9same name
29 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository56 other files it gives its agents.
- autopilotSkill
- before-afterSkill
- bigpictureSkill
- blast-radiusSkill
- brainstormSkill
- compoundSkill
- confidence-checkSkill
- contextualizeSkill
- create-ship-skillSkill
- create-skillSkill
- curateSkill
- debugSkill
- deep-researchSkill
- devtoolsSkill
- distillSkill
- done-checkSkill
- doneSkill
- double-checkSkill
- envsetSkill
- explainSkill
- fortifySkill
- goSkill
- handoffSkill
- harness-auditSkill
- interviewSkill
- matrixSkill
- mockupSkill
- mutateSkill
- notesSkill
- postmortemSkill
- presentSkill
- qaSkill
- readableSkill
- recommendSkill
- reportSkill
- retireSkill
- review-planSkill
- scoutSkill
- sec-reviewSkill
- setup-ticketingSkill
- setup-worktreesSkill
- spec-checkSkill
- specSkill
- stepsSkill
- successionSkill
- super-reviewSkill
- tangentSkill
- tddSkill
- tend-memorySkill
- testSkill
- ticketSkill
- triageSkill
- waitSkill
- workSkill
- worth-itSkill
- wrap-upSkill
The file
We split this skill into the rules it gives an agent
6 rules. 0 are common, 6 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 34 edited versions we read keep these.
6
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
- Distinctive: line 14
If I read only this line, I know where things stand.
- Distinctive: line 44
ℹ️ is the fallback, not the filler: Never force a bullet into a status it doesn't have — a wrong icon is worse than a neutral one.
- Distinctive: line 59
If naming it would produce filler — Note, Thing, Update — or if the sentence *is* the point, skip the label and let the line run from the …
- Distinctive: line 62
Don't also write the status in words — "✅ Guards work: tested live" not "✅ Done — guards work".
- Distinctive: line 65
No preamble ("Here's the TL;DR…"), no closing wrap-up.
- Distinctive: line 104
If I've added a focus below, narrow to just that slice: $ARGUMENTS
“Common” is measured: kept by at least 6 in 10 of the 34 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.
Files
Give me the TL;DR of your last response. High signal, low noise — compress, don't re-explain. Assume I'm cognitively loaded or short on time.
Shape
1. Bottom line — one sentence, first thing, no icon. The decision, outcome, or key finding. If I read only this line, I know where things stand.
2. Icon-keyed bullets — flat, one line each, no nesting. Anatomy:
- <icon> **<label>** — <the line>
│ │ └─ compressed, one line, no nesting
│ └─ optional: one or two words naming the subject
└─ required: exactly one, from the key belowThe icon is the primary sorting device: scanning the left edge alone tells me which lines are settled, which are live, which will bite me. The label is the second axis — it lets me find the lunch thing or the timer thing without reading cells.
Cut reasoning, caveats, educational asides, and restatements of my question. Keep only what I need to know or remember.
Icon key
| ✅ | good / done / verified / works as intended |
| ⏳ | in flight, pending, or waiting on something — a running job, an open question with someone else, a timer that hasn't fired |
| ⬜ | outstanding — stated in the response as not-yet-done |
| ⚠️ | trap, gotcha, or constraint that will bite if ignored |
| ❌ | failed, broken, ruled out, or won't work |
| ❓ | unverified / unknown / assumed but untested |
| ℹ️ | plain context — a fact, a value, a design choice with no status attached |
ℹ️ is the fallback, not the filler. Never force a bullet into a status it doesn't have — a wrong icon is worse than a neutral one. If most bullets land on ℹ️, that's the honest read: the response was informational.
Rules
- Icons classify what was already said. They never add anything. A ⬜ marks a fact of state
the response stated ("not yet in the prod config"), not a task I'm being handed. This skill
still emits no Do-next block, no action plan, no recommendation — chain
/recommendif that's what I want. - Keep the response's order. The icons do the sorting; reordering breaks the thread. One exception: a ⚠️ or ❌ that changes the bottom line gets pulled to the top and shows up in the bottom-line sentence.
- One icon per bullet. Never two, never a bullet without one.
- Label only when the bullet has an obvious subject. One or two words, bolded, em-dash after. If naming it would produce filler — Note, Thing, Update — or if the sentence is the point, skip the label and let the line run from the icon. A mix of labelled and bare bullets in one list is fine and normal; forcing every line to carry one is not.
- Don't also write the status in words — "✅ Guards work: tested live" not "✅ Done — guards work".
- Trailing bare facts that belong to no bullet (a path, a command) can sit on their own line after the list, unadorned.
- No preamble ("Here's the TL;DR…"), no closing wrap-up.
Worked example
**Cache layer shipped behind a flag; only the prod config flip remains.**
- ✅ **Read path** — cache hits verified locally; p95 dropped from ~420ms to ~60ms
- ✅ **Invalidation** — writes evict the key in the same transaction, covered by two new tests
- ⚠️ **Stale reads** — the admin screen bypasses the write path, so its edits stay stale for
up to the 5-minute TTL
- ⏳ **Staging deploy** — pipeline still running at time of writing
- ℹ️ **TTL** — 5 minutes, chosen to match the upstream data refresh cadence
- ℹ️ Kept the existing client library — adding a second one for one feature wasn't worth it
- ❓ **Memory** — cache size under real traffic is unmeasured; estimate assumes ~10k keys
- ⬜ **Prod flag** — `CACHE_ENABLED` is still off in prod config
src/cache/config.tsNote the client-library line carries no label — a label would just repeat the first words of the sentence. That's the rule working, not a miss.
Not a table
Deliberately bullets, not a table. A two- or three-column markdown table scans slightly better on a wide terminal, and loses on a narrow one:
- The CLI demotes the whole table to a borderless stacked fallback as soon as any single
cell wraps past four lines. On a narrow terminal (a phone SSH client, ~40 cols) a TL;DR line —
compressed prose, not a status fragment — routinely blows that, so the table collapses exactly
where a quick read matters most.
/reportsurvives it only because its cells are terse and budgeted to ~320 chars. /reportalready is the icon table. A second one makes two different jobs — where-things- stand vs. what-you-just-said — indistinguishable in scrollback.- The icon already sits at a fixed column in bullet form, so the table's alignment win is nearly nil. The label column was the real gain, and the bold label above delivers it with no width cliff.
If I've added a focus below, narrow to just that slice: $ARGUMENTS
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
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.

