minimize
Skill1 starsChanged 2 months agoMITView on GitHub
Give a short, chill rundown of what's actually happening across the user's projects — recent work, live blockers, what's shipped, what's owed. Cuts the tech noise into a scannable digest. Use when the user invokes /minimize, or says "minimize", "what's going on", "catch me up", "where are we", "status", "what's the state of things", "sitrep", or "give me the rundown".
In this repository7 other files it gives its agents.
- artifact-to-pdfSkill
- character-sheet-gateSkill
- chill-dudeSkill
- code-rendered-videoSkill
- coding-cli-brief-loopSkill
- deploy-and-verifySkill
- emojliSkill
The file
We split this skill into the rules it gives an agent
8 rules. About 0 are common, 8 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈8
Distinctive · estimate
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 12
Signal only: If it wouldn't change what they do next, cut it.
- Distinctive: line 22
Don't over-gather. Two or three sources is plenty.
- Distinctive: line 26
Keep it tight. Structure like this, but only include lines that have something:
- Distinctive: line 48
If everything's calm, say so in one line and stop.
- Distinctive: line 56
Don't re-explain project history they already know.
- Distinctive: line 57
Don't list every commit or every observation.
- Distinctive: line 58
Don't editorialize or add "next steps" unless they're genuinely blocking.
- Distinctive: line 59
Don't end with "let me know if you'd like more detail" — just leave the door open by being brief.
≈ “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.
Files
Minimize 🌊
Give the user a chill, scannable picture of where things stand. Not an essay. Not a report. Just: here's the vibe, here's what's hot, here's what's owed. Then stop.
The one rule
Signal only. If it wouldn't change what they do next, cut it. The whole thing should read in ~20 seconds.
Gather (fast, parallel)
Pull from whatever's available — run these together, don't block on any one:
- claude-mem — most recent observations. The session-start
$CMEMblock is often already in context; use it. Otherwise callmcp__plugin_claude-mem_mcp-search__timelineorsearchfor the last day or two. - Live blockers — if a project-tracking MCP is connected, pull its open blockers/risks. Skip silently if not.
- Git — for the 2-3 repos touched most recently, a quick
git -C <path> log --oneline -5andgit -C <path> status -s. Uncommitted work and un-pushed branches are the real "what's going on."
Don't over-gather. Two or three sources is plenty. If one's slow or missing, drop it and move on — never make the user wait.
Format the rundown
Keep it tight. Structure like this, but only include lines that have something:
🌊 Here's the vibe — {date}
🔥 Hot right now
• {the 1-3 things actively in motion}
✅ Landed
• {what shipped since last check — with the win, not the mechanics}
⛔ Owed / blocked
• {what's waiting on you or stuck — the ONE thing that matters most first}
😎 All quiet elsewhere.Rules:
- Max ~3 bullets per section. More than that = you're dumping, not curating.
- Lead each bullet with the outcome, not the plumbing. "Sharing is live in prod" beats "merged PR #14 into main at f3ad4c9."
- Drop empty sections entirely. No "Nothing to report here" filler.
- One emoji per section header. Don't carpet-bomb.
- If everything's calm, say so in one line and stop. That's a valid answer.
Tone
Laid-back but sharp — think a good teammate leaning over your desk, not a Jira export. Plain sentences. If the user has chill-dude vibes going, match it; otherwise stay warm and terse. Never pad. Never apologize for a short answer — short is the answer.
Don't
- Don't re-explain project history they already know.
- Don't list every commit or every observation.
- Don't editorialize or add "next steps" unless they're genuinely blocking.
- Don't end with "let me know if you'd like more detail" — just leave the door open by being brief.
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.

