tighten
Skill4 starsChanged 8 days agoMITView on GitHub
Condense verbose prose in project docs without dropping facts or restructuring; docs only, never emails or messages. Use when the user says "tighten this doc", "condense these files" or "make this less wordy".
10 other projects use a version of this.
Adopted by
Most recently taken up by
- aditya-vermax/skillstheir own version, first seen Nov 8, 2026
- nearmap/shipyardtheir own version, first seen Nov 8, 2026
- nuko-nova-dynamics/nuko-nova-unsloptheir own version, first seen Nov 6, 2026
- wtsi-hgi/agentskillstheir own version, first seen Nov 1, 2026
- harshil-1411/claude_plugin_video_studiotheir own version, first seen Oct 17, 2026
- captkernel/claude-writing-toolstheir own version, first seen Oct 15, 2026
Other versions10 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 10
- autobrr/mkbrr★ 363same name
- aditya-vermax/skills★ 1same name
- corygabrielsen/skills★ 1same name
- markrogersjr/skills★ 1same name
- nearmap/shipyard★ 1same name
- nuko-nova-dynamics/nuko-nova-unslop★ 1same name
- wtsi-hgi/agentskills★ 1same name
- captkernel/claude-writing-tools★ 0same name
- harshil-1411/claude_plugin_video_studio★ 0same name
- raj-khan/tighten★ 0same name
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 repository15 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- cavemanSkill
- gh-issues-by-severitySkill
- grid-helpSkill
- grill-meSkill
- growth-psychSkill
- handoffSkill
- mine-learningsSkill
- openspec-helpSkill
- rubber-duckSkill
- setup-repo-skillsSkill
- skill-scoutSkill
- spec-scoutSkill
- standupSkill
- un-claudishSkill
The file
We split this skill into the rules it gives an agent
2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 8 edited versions we read keep these.
2
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
Never assume a default scope (not "whole project," not "last-touched files").
- Distinctive: line 71
Do not commit. Tell the user the diffs are ready for git diff review — a multi-file rewrite is worth eyeballing before it lands.
“Common” is measured: kept by at least 6 in 10 of the 8 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
tighten
Rewrite existing prose in files to be shorter and higher-signal, in place. This is a density pass, not a rewrite — meaning, facts, and structure must survive exactly.
Scope
Never assume a default scope (not "whole project," not "last-touched files"). Always ask for the file list/glob if the invocation didn't include one, every run — scope is not remembered between runs.
Once you have candidate files, auto-drop these before confirming (no need to ask about them individually):
- Any file whose header/frontmatter marks it generated or do-not-edit-by-hand (e.g. "generated by X — do not edit", "Never edit by hand").
- Known generated indexes in this repo, e.g.
SKILLS.md.
Confirm the final resolved file list back to the user before editing anything. If it's empty after exclusions, say so and stop.
Style target
Docs only. Never apply email/message conventions — no shorthand ("pls", "thx"), no sign-off, no informal tone shift.
Load these sources each run, when available:
- A personal email-style reference, if you keep one (e.g. in a notes vault
under something like
<vault>/email-style.md) — use only the conciseness principles (the core "if it can be said in one sentence, do that" rule, short sentences, bullets over prose, no fluff/filler/jargon). Ignore its VOICE section's informalities and skip the ENDING section entirely — those are email-specific. If this file is missing or unreadable, skip it and note that in the final report. - A communication-style rules file (banned jargon words and banned filler phrases), only if the user names one — use it as the hygiene checklist. If none is named, skip it and note that in the final report.
Core test per sentence: "if it can be said in one sentence, do that." Cut filler, hedging, restated ideas, throat-clearing preamble, and trailing summaries. Prefer bullets to prose.
What not to touch
- Code blocks, frontmatter, tables/data, links, headings/structure.
- Facts, numbers, caveats, conditionals — meaning must survive exactly.
- Section order and heading structure — this is a density pass, not a reorganization.
How to run
- Resolve scope (above) and confirm the file list with the user.
- Spawn one agent per file, in parallel (files are independent — no shared
state, so no need to serialize). Each agent gets: the file path, the
style-target excerpt from above, the "what not to touch" constraints, and
instructions to:
- Read the file.
- Rewrite prose sections tighter, editing in place.
- Report back: word count before, word count after, one-line summary of what changed. If the file had no cuttable prose (already tight, or all code/tables), report 0% change rather than forcing an edit.
- Collect all reports into one summary table: file | before | after | % reduction | note.
- Do not commit. Tell the user the diffs are ready for
git diffreview — a multi-file rewrite is worth eyeballing before it lands.
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.

