All skills
simota avatar

/crest

@8e1f365
by shingo imotasimota/agent-skills85 stars
15

Building engineer self-branding by turning technical contributions into a professional brand. Use for GitHub/LinkedIn/blog/conference positioning or content strategy.

Use this Skill: https://skilld.dev/gh/simota/agent-skills/crest

This session only. Nothing lands on disk.

referencetopic-dna.md

≈3k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Topic DNA

Purpose

"What are you known for?" is the brand question every engineer eventually faces. Topic DNA — the intersection of Tech × Domain × Perspective — answers it. A defined Topic DNA gives every profile, post, talk, and interview a unifying axis. Without one, the engineer is forgettable in a crowded market.

Scope Boundary

  • IN scope: positioning statement authoring, Tech × Domain × Perspective triangulation, content-pillar derivation, defensibility audit, durability check, audience-fit verification.
  • OUT of scope: SNS publishing strategy (sns), content calendar / amplification (blog), profile copy on a specific platform (github / linkedin / bio), portfolio architecture (portfolio), full-funnel marketing (delegate to growth / funnel).

Core Concepts

The Triangulation

A defensible niche sits at the intersection of three axes:

Axis Question Examples
Tech What do you build with? Rust, TypeScript, Postgres, Kubernetes, ML, Game engine
Domain What domain do you build for? Fintech, biotech, devtools, climate, music, accessibility
Perspective What is your distinctive angle? Performance obsessive, accessibility-first, indie-hacker, security-first, teaching-driven, OSS maintainer

A pure-Tech identity ("Rust engineer") is replaceable. A pure-Domain identity ("fintech engineer") is generic. A pure-Perspective identity ("performance obsessive") is too vague. The Tech × Domain × Perspective intersection is where credibility compounds.

Examples of Crisp Topic DNA

Engineer Tech Domain Perspective One-Sentence Topic DNA
Linus Torvalds C, kernel Operating systems Pragmatic curmudgeon "Open-source kernel maintainer with a sharp tongue and zero patience for stupidity."
Sara Soueidan HTML, CSS, SVG Web Accessibility-first "SVG and accessibility expert teaching the web to be more humane."
Brad Frost HTML, CSS, design-system tooling Web Atomic systems "Design-system advocate with a method for building scalable component libraries."
Julia Evans Linux, networking Devtools Teaching through cartoons "Systems engineer who turns hard topics into illustrated zines."
Kelsey Hightower Go, Kubernetes Cloud-native Teach-by-living "Cloud-native engineer who runs production from a Raspberry Pi to prove a point."
Cassidy Williams JavaScript, devtools Developer experience Humorous + earnest "DX advocate who memes the JS ecosystem with weekly newsletters and conference talks."
Wes Bos JavaScript, web Education Tutorial-driven "JS teacher producing premium courses on modern web stacks."

Crisp Topic DNA is identifiable in one sentence and survives a 5-second elevator pitch.

Defensibility Audit

A Topic DNA must be:

Test Question
Specific Can someone repeat it back accurately after one read?
Defensible Are there fewer than 100 engineers globally who hold this exact intersection?
Demonstrable Can you point to 3+ public artifacts (repos / posts / talks) that prove it?
Audible When others describe you, do they use words from your Topic DNA?
Durable Will this still be relevant in 12 months? In 3 years?
Genuine Do you actually want to do this work, or are you chasing a trend?

If all six pass, the Topic DNA is shipping-ready.

Durability vs Trend

Avoid Topic DNA tied to single technologies that age fast:

Risk Trend Durable Reframe
High "Web3 NFT engineer" (2021) "Distributed systems with cryptographic verification"
High "Prompt engineering specialist" (2024) "AI evaluation and reliability for production systems"
Medium "Tailwind CSS expert" "CSS architecture for maintainable design systems"
Low "TypeScript performance optimization" (the work and language are stable)

Tech component should be a class of problem, not a hype tool.

Content Pillars from Topic DNA

Once Topic DNA is set, content automatically derives:

Topic DNA Pillar 1 Pillar 2 Pillar 3 Pillar 4
"Postgres performance for fintech" Index design Query plans Replication / HA Compliance & audit
"TypeScript DX in monorepos" Type system depth Tooling (ts-config, build) Editor / LSP integration Migration patterns
"Accessibility-first React" ARIA patterns Keyboard / focus Screen-reader testing Color & contrast tooling

Aim for 3–5 pillars. Each pillar generates years of articles, talks, and projects.

The 80/20 Rule

LinkedIn's 360Brew Topic DNA model recommends:

  • 80% of content within Topic DNA pillars (depth signal).
  • 20% outside the pillars (breadth signal — keeps you human).

Below 80% on-pillar dilutes positioning. Above 95% becomes one-note.

Audience Fit

Topic DNA must intersect a real audience:

Audience size Signal
< 1,000 globally Niche of niche; possibly too narrow unless fee per client is high
1,000–10,000 Healthy expert niche
10,000–100,000 Strong niche; can sustain courses, books, conferences
100,000+ Mass-market expertise; harder to differentiate

Smaller niches monetize per-client at higher rates; larger niches at scale via products. Pick based on monetization preference.

Common Failure Modes

Failure Symptom Fix
Buzzword soup "AI-powered cloud-native serverless TypeScript engineer" Pick one Tech, one Domain, one Perspective
Pure resume "Senior software engineer with 10 years experience" Move from history to current obsession
Trend-tied "AI agent prompt engineer" Reframe to durable skill class
Brand-dependent "Ex-Google engineer" Insufficient — what do you DO?
Vague perspective "Engineer who cares about quality" What kind of quality? performance? a11y? security?
Over-broad "Full-stack engineer" Too generic for positioning; reserve for skills section
Aspirational "Future founder building the next Stripe" Build the work first, brand the work second

Triangulation Worksheet

A practical exercise:

  1. List 5 technologies you've shipped non-trivial work in (last 24 months).
  2. List 3 domains you've worked in or have credible interest in.
  3. List 5 perspectives that describe how you approach problems differently.
  4. Form 5 candidate triangulations: Tech[i] × Domain[j] × Perspective[k].
  5. Run defensibility audit on each.
  6. Pick top 1–2 and write the one-sentence positioning.

Two-Tier Strategy

Some engineers maintain:

  • Primary Topic DNA — used in headline, About, conference bios.
  • Secondary Topic DNA — used for side-projects, experiments, hobby.

Two-tier is fine if both are clearly demarcated. The risk is signal dilution if both appear at primary positions.

Evolution Cadence

Topic DNA should evolve every 18–36 months, not quarterly. Stability builds credibility; constant rebranding signals confusion.

Trigger to update Healthy
Major career change (role, industry) yes
Public artifact corpus shifted significantly yes
Audience grew beyond original niche yes
You stopped enjoying the work yes
Trend is fading yes (re-anchor on durable skill)
New tool dropped this week no
Quarterly "rebrand" cycle no

Workflow

  1. List 5 tech axes / 3 domains / 5 perspectives from real shipped work.
  2. Form 5 candidate triangulations.
  3. Draft a one-sentence positioning per candidate.
  4. Run defensibility audit — 6 checks (specific / defensible / demonstrable / audible / durable / genuine).
  5. Audience-fit check — niche size estimate.
  6. Derive 3–5 content pillars from the chosen Topic DNA.
  7. Apply 80/20 ratio plan for content allocation.
  8. Document Topic DNA in a single canonical artifact (used to seed bio, profile, talks).
  9. Set evolution review for 18–36 months out.
  10. Hand off to bio for platform variants, portfolio for site IA, blog / sns / conference for content amplification.

Output Template

topic_dna:
  primary:
    one_sentence: "Postgres performance specialist for fintech systems with a compliance-first lens."
    tech: postgres
    domain: fintech
    perspective: compliance_first
    defensibility:
      specific: PASS
      defensible: PASS    # ~50 globally
      demonstrable: PASS  # 3 OSS contributions, 4 articles, 2 talks
      audible: PASS       # peers describe me with these terms
      durable: PASS       # 3-year horizon
      genuine: PASS       # genuinely interested
    audience_estimate:
      size_band: 5000_to_15000
    content_pillars:
      - id: P1
        name: "index design and bloat management"
        share_pct: 25
      - id: P2
        name: "query plan diagnostics under load"
        share_pct: 25
      - id: P3
        name: "replication, HA, and audit logging"
        share_pct: 20
      - id: P4
        name: "compliance, regulatory query patterns (PCI/SOX/MAS)"
        share_pct: 20
      - id: P5
        name: "off-pillar (cooking, climbing, side bets)"
        share_pct: 10
    on_pillar_target_pct: 80
    review_due: 2027-10-01
  secondary:
    enabled: no
  handoff:
    bio: derive_platform_variants
    portfolio: site_ia
    blog: pillar_calendar
    sns: zero_click_publishing

Anti-Patterns

  • Buzzword soup — combining 4+ technologies in the headline.
  • Trend-bound Tech axis — dies with the trend.
  • Pure resume framing — past, not present.
  • Vague Perspective ("quality-focused", "team player").
  • Topic DNA that doesn't match what you actually do daily.
  • Aspirational DNA before any artifacts back it up.
  • Quarterly rebranding — no compounding.
  • Two-tier with unclear demarcation — confuses everyone.
  • Niche under 1,000 globally without high fee-per-client model.
  • Niche over 100,000 with no differentiation — invisible.
  • Topic DNA derived from competitor envy rather than genuine interest.
  • Topic DNA that requires explanation longer than one sentence.
  • Pillars overlapping — same content categorized two ways.
  • Pillars below 3 (too narrow) or above 5 (too scattered).
  • 100% on-pillar content — robotic, no humanity signal.
  • Below 70% on-pillar — positioning collapses.

Deliverable Contract

A Topic DNA package is complete when:

  • One-sentence positioning written.
  • Tech / Domain / Perspective named explicitly.
  • Defensibility audit passes 6/6 checks.
  • Audience size band estimated.
  • 3–5 content pillars derived with share % summing to 100.
  • 80/20 on-pillar policy stated.
  • Review date set 18–36 months out.
  • Handoff briefs prepared for bio, portfolio, blog, sns, conference.

References

  • Marty Neumeier, Zag (2007) — radical differentiation framework.
  • Al Ries & Jack Trout, Positioning (1981) — original positioning theory.
  • April Dunford, Obviously Awesome (2019) — positioning for technical products (transferable to engineers).
  • Patrick McKenzie, Don't Call Yourself a Programmer — engineer self-marketing primer.
  • David Heinemeier Hansson, Rework / REMOTE — opinionated engineer brand examples.
  • LinkedIn 360Brew — Topic DNA model for the platform.
  • Jason Fried, It Doesn't Have to Be Crazy at Work — durable, calm brand voice.
  • Cal Newport, So Good They Can't Ignore You (2012) — career-capital framing.
  • Daniel Pink, Drive — autonomy-mastery-purpose alignment with brand work.
  • Seth Godin, This Is Marketing (2018) — niche-first audience strategy.
  • Justin Welsh, The Diversified Solopreneur — one-niche solo creator model.
  • Andy Crestodina, Orbit Media — content-pillar methodology.

Source: SKILL.md on GitHub

1 warning4mo4 checks · Risk SAFE
  • Gen Agent Trust Hub4mo

    The 'crest' skill is a comprehensive engineer self-branding strategist designed to help users transform technical contributions into professional narratives across platforms like GitHub, LinkedIn, and personal blogs. It focuses entirely on content strategy, positioning, and profile optimization, with no executable code or dangerous capabilities. The skill emphasizes authenticity, data-backed achievements, and adherence to professional anti-patterns.

  • Socket4mo

    No alerts

  • Snyk4mo

    Risk: MEDIUM · 1 issue

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at 8e1f365. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 3 days ago.

Activeupdated last month

README badge

README badge for simota/agent-skills/crest