writing-technical
Skill1 starsChanged 3 months agoMITView on GitHub
Technical writing form — multi-mode structures for findings, field logs, claim diaries, bake-offs, design briefs, concept maps, and forum notes. Mechanism-first, evidence objects, reframe insights, open gaps. Domain-agnostic (not only data/math). Prefer writing-docs for pure procedures; writing-prose for general human tone; writer-style + evan pack for full Evan cadence and mode routing; writing hub when unclear.
In this repository47 other files it gives its agents.
- data-apache-lakehouseSkill
- data-apiSkill
- dataSkill
- data-duckdbSkill
- data-identity-resolutionSkill
- data-pipeline-operationsSkill
- data-pokemon-tcg-foundationsSkill
- data-product-evalSkill
- data-reelsSkill
- data-semantic-qualitySkill
- data-table-lifecycleSkill
- html-designSkill
- ebay-listingSkill
- flutter-pro-ux-reviewSkill
- flutterSkill
- browser-testing-with-devtoolsSkill
- browser-verifySkill
- design-systemSkill
- frontend-designSkill
- mobile-app-echarts-visualSkill
- mobile-blog-chart-visualSkill
- mobile-chart-visualSkill
- mobile-product-uxSkill
- mockup-implementSkill
- product-designSkill
- product-ui-craftSkill
- prototypeSkill
- react-performanceSkill
- ui-exploreSkill
- visual-verifySkill
- web-qualitySkill
- high-retention-insight-reels-instagramSkill
- instagramSkill
- instagram-virality-insightsSkill
- marketing-cashvertisingSkill
- marketing-contagiousSkill
- marketing-going-viralSkill
- marketingSkill
- marketing-offersSkill
- marketing-storybrandSkill
- quality-checkSkill
- elements-of-styleSkill
- style-clarity-graceSkill
- writer-styleSkill
- writing-docsSkill
- writing-proseSkill
- writingSkill
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 44
Default if unclear: findings-note.
- Distinctive: line 63
Prefer incentive/structural implication, “slogan incomplete,” or “property falls out of the map” over mood (“concerning,” “exciting”).
- Distinctive: line 76
If you close: implication or next experiment.
≈ “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
Writing technical — form by job (multi-mode)
Job: Explain a system, result, design, or learning path so a technical reader can reproduce the reasoning (and often the measurement or the map).
Aligns with Evan Kim’s public archive presentation modes — same mind, different
page shapes. Full mode cards: writer-style/profiles/evan/modes.md.
When to load
- Technical articles that teach, report, compare, or design
- Empirical notes, build logs, paper takeaways, tool bake-offs, design briefs
- Any domain (infra, markets, math, agents, product internals)
Not for: pure runbooks (writing-docs), landing pages (marketing), persona cosplay without a form (writer-style alone after form is set).
Step 0 — Pick a mode (required)
| Job | Mode | Skeleton |
| Report a measured/observed result | findings-note | TL;DR → setup → results → reframe → open Q → appendix |
| What production taught you | field-log | Constraint → built → problem/fix/lesson → next |
| Digest a paper/theory in public | claim-diary | Frame → numbered claims-as-H2 → quote/gloss → open gaps |
| Argue design A vs B | systems-essay | Landscape → mech A → mech B → costed pick |
| Same task, multiple tools | bake-off | Task → parallel sections per candidate → differential |
| Narrative how-to with a visible result | walkthrough | Goal → steps + artifacts → what you should see → appendix |
| Pure install/runbook | → writing-docs | — |
| New lens on a structure | concept-map | Representation → worked example → properties → questions |
| Artifact + why that shape | design-brief | TL;DR bounds → constraints → technique → applications |
| Partial public research | forum-fragment | Observation → partial implication → Open questions |
Default if unclear: findings-note.
Wrong mode > slightly off tone.
Principles (all modes)
1. Structure does the thinking
Outline is the argument. Headings advance claims; parallel sections enable comparison.
2. Payload early
Finding, constraint, thesis, claim list, or representation rule in the first screen.
3. Evidence objects
Number+unit, quote, table, chart, query, error string, diagram, config — next to the claim.
4. Insight as reframe
Prefer incentive/structural implication, “slogan incomplete,” or “property falls out of the map” over mood (“concerning,” “exciting”).
5. Facts first
Fact-sheet for every quantitative or citation-sensitive claim; style after freeze. Cross-check when two sources exist.
6. Open gaps stay open
Unproven bridges and next experiments are first-class — especially claim-diary and forum-fragment.
7. Conclusion discipline
If you close: implication or next experiment. Ban “In conclusion, we have shown…” restatements.
8. Voice handoff
House technical (this skill) or writer-style pack evan (loads modes.md + exemplars for the chosen mode).
Anti-patterns
- Routing form by topic (“math = academic essay”) instead of job
- Field-log scar cosplay on a paper diary
- TL;DR theater with no findings
- Academic filler connectives without new claims
- Charts without stating what changed
- Homogeneous sentence polish (AI default)
Hand off
| Need | Skill |
| Install/runbook only | writing-docs |
| Full Evan cadence + mode exemplars | writer-style (evan) |
| Deslop only | writing-prose |
| Turgid syntax | style-clarity-grace |
| Usage / concision mechanics | elements-of-style |
| Offer/landing | marketing |
Done criteria
- Mode named and skeleton followed
- Payload in first screen
- Evidence objects on load-bearing claims
- Reframe or open gap — not padded summary
- Optional: evan voice pack if “sound like Evan” requested
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.