tighten-instruction
Skill4 starsChanged 7 months agoMITView on GitHub
Make an instruction line read cold: clarify it into plain words, then tighten it to one positive line. TRIGGER when: user says 'tighten/clarify this rule/fact', 'make this leaner/simpler/clearer', 'titan instruction' in a skill, CLAUDE.md, agent prompt, or rule file.
In this repository42 other files it gives its agents.
- AGENTS.mdAGENTS.md
- agent-soulSkill
- audit-transcripts-for-learningsSkill
- best-answerSkill
- check-coherenceSkill
- compress-fileSkill
- deep-modulesSkill
- design-critiqueSkill
- durable-docs-updateSkill
- execute-chatSkill
- execute-planSkill
- explain-deeplySkill
- explain-simplySkill
- find-gapsSkill
- fix-verify-loopSkill
- generate-mockupsSkill
- grill-meSkill
- handoffSkill
- interview-meSkill
- jtbdSkill
- map-design-languageSkill
- memory-pruneSkill
- multi-agent-analysisSkill
- place-factSkill
- post-purchase-ui-extensionSkill
- product-interviewSkill
- prove-behaviorSkill
- refine-fileSkill
- second-opinionSkill
- sentry-analysisSkill
- shopify-dev-mcpSkill
- structure-proseSkill
- tech-designSkill
- test-completed-planSkill
- tighten-fileSkill
- tldrSkill
- triageSkill
- trim-specSkill
- two-pass-reviewSkill
- validate-answerSkill
- vet-factSkill
- write-planSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
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 23
Before: *"Run tests after edits.
- Distinctive: line 24
After: *"Run tests after every edit."*
- Distinctive: line 29
Test it cold: Read the final line without surrounding context — every phrase obvious, the whole line actionable.
≈ “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
Tighten Instruction
Two passes in order: clarify the text into plain words, then tighten each distinct instruction or fact to one line — clarity first, so tightening never compresses back into jargon.
Steps
- State the goal in one sentence: what should the reader do or know after reading it?
Clarify — make every phrase read cold (clear on its own). Decompose the line into phrases; name each phrase's intended meaning and fix any that won't land:
- Jargon → plain word ("down-weight" → "rank lower").
- Abstraction → concrete verb/noun ("optimize for" → "rank by").
- Ambiguous referent → name it ("them" / "the better one" → the actual subject).
- Mismatched verb → the true actor ("pick" when you only recommend → "favor").
- Cryptic compression → unpack it ("not least effort" → "not the easiest option").
Spend length here when it buys comprehension; preserve the meaning exactly.
Tighten — cut what the clear line doesn't need. Drop any line whose only purpose is "restate the goal," "hedge," or "explain why"; keep a non-derivable rationale as a fact in Step 4. Collapse clauses the positive form already implies, and rename a heading that misnames its content. Never trade a plain word back for jargon to save space.
Before: "Run tests after edits. Don't skip even for small changes — small changes break things too." After: "Run tests after every edit."
- Land each distinct instruction or fact on one positive line — in the shape the content wants:
- Instruction → trigger + action: "When X, do Y." / "Use X for Y."
- Fact → subject + the non-derivable part: "X is Y, not the expected Z." / "X — the constraint/reason." (gotchas, couplings, conventions, rationale all take this shape)
- If it won't compress to one line: split — it's two.
- Test it cold. Read the final line without surrounding context — every phrase obvious, the whole line actionable. If not, redo.
- Score and gate the edit. Score confidence
0.00–1.00that the final line is clearer, tighter, and meaning-preserving. If a caller owns a file-level gate, return the score and proposal without editing. Otherwise, apply atc ≥ 0.80; below that, keep the current line and report the proposal as held.
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.

