agentleFS
Ethan9123Ethan9123/wx-life-analysis/.claude/skills/mbti-analysis/SKILL.md

mbti-analysis

Skill0 starsChanged 6 months agoApache-2.0View on GitHub

Infer a WeChat contact's MBTI type from their chat history AND their Moments / SNS feed combined, identify trip wires (topics that shut them down), surface persona splits (public-extrovert vs private-reserved), and translate everything to a concrete comms strategy (frequency, style, do-list, avoid-list). Activate when the user wants to understand what type a contact is, why a conversation cooled, what's safe to bring up, or how often to message them. Reads BOTH people/<slug>/chat.md (private 1:1, ~70% signal) AND people/<slug>/sns.json (public self-presentation, ~30% signal, often flips borderline axes). Writes to people/<slug>/profile.md YAML frontmatter under mbti / trip-wires / comms fields plus a "朋友圈观察" body section. Uses observable signals only — reply latency, voice/text ratio, abstract vs concrete language, planning style, SNS post frequency, caption tone, audience scope, visible gaps — never fabricates certainty (always records a confidence level).

2 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions2 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 2

    1. anneheartrecord/charles-skill★ 3same name
    2. anneheartrecord/talk-mbti★ 2same 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 repository7 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 1 is distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    When this skill is loaded, you are running a per-person personality + comms-strategy analysis over a contact's WeChat chat history.

    line 8
  2. Very specific:

    Full methodology: read docs/mbti-analysis.md before producing output.

    line 10
  3. Very specific:

    If SNS shows a polished extrovert but chat shows reserved/tired — don't approach with the public version of them.

    line 39
  4. Very specific:

    Always record mbti.confidence: low | medium | high plus 2-4 sentences of mbti.basis citing specific data points, not vibes.

    line 47
  5. Very specific:

    Don't paste chat.md excerpts outside the local session: All evidence stays in the user's local clone.

    line 75

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

MBTI Analysis Skill

When this skill is loaded, you are running a per-person personality + comms-strategy analysis over a contact's WeChat chat history.

Full methodology: read docs/mbti-analysis.md before producing output. The methodology document defines the 4-axis scoring table, the trip-wire taxonomy, and the comms strategy translation rules.

Quick reference (don't substitute for reading the methodology)

Input sources (read BOTH)

  1. people/<slug>/chat.md — private 1:1 history. ~70% of useful signal. Strongest for trip wires + comms style.
  2. people/<slug>/sns.json — public Moments feed. ~30% of useful signal. Often flips a borderline axis call. Critical for surfacing persona splits and safe/unsafe topic territory.

Four MBTI axes — what to look for (chat signals)

AxisSkew toward first letterSkew toward second letter
E/Ihigh message volume, fast replies, voice notes, skims topicslower volume, batched replies, prefers text, depth dives
S/N"where/when/how-much" questions, concrete nouns, present-focused"why/what-if" questions, abstract concepts, future-focused
T/Flogic-first, blunt disagreement, "smart take" praiseempathy-first, cushioned disagreement, "thoughtful" praise
J/Pproposes specific time+place, locked plans, lists"free this week, you pick", flexible, last-minute OK

Four MBTI axes — what to look for (SNS signals)

AxisSkew toward first letterSkew toward second letter
E/Iposts 3+/week, comments on others, broad audienceposts <1/week, mostly read-only, "close friends" audience
S/Nconcrete captions ("今天在 X 吃了 Y"), documentary photos, practical share-linksabstract captions, aesthetic-curated photos, philosophical share-links
T/Fachievements / news / sports, sparse emotional disclosuremood posts, named feelings, "for me this means…" framing
J/Pplanned recaps ("年终") / curated trip albums / themed collectionsspontaneous moments, unedited, jumps topic

Persona splits

If SNS shows a polished extrovert but chat shows reserved/tired — don't approach with the public version of them. Note explicitly in mbti.basis. This is one of the highest-value observations the SNS pass produces.

Voice transcription (optional depth-add)

For voice-heavy contacts (voice/text ratio > 5%), the agent should suggest running tools/voice-transcribe.ps1 / .sh to surface content from voice messages. Read the resulting people/<slug>/voice-transcripts.md alongside chat.md — voice often carries the emotional/vulnerable register that text doesn't. Methodology: docs/voice-transcription.md. Privacy default is local-only Whisper; cloud APIs require explicit user I CONSENT confirmation.

Confidence required

Always record mbti.confidence: low | medium | high plus 2-4 sentences of mbti.basis citing specific data points, not vibes. If signals are mixed, write "INFP/ENFP" or "unclear".

Trip wires — what shuts them down

Walk the chat for: latency spikes (3-5× baseline), length collapse, hard topic redirect, soft non-answer, "you sound like a bot" meta-critique. Need 2+ signals from different categories before calling something a trip wire — single signals are noise.

Output schema

mbti:
  type: ENFP                  # or "INFP/ENFP" if borderline, or "unclear"
  confidence: medium          # low / medium / high
  basis: |
    <2-4 sentences citing specific signals from chat.md>
trip-wires:
  - topic: <short label>
    pattern: <what triggered → reaction>
    repair: <how to avoid / recover>
comms:
  frequency: <e.g. "1 ping / 2 days when ball mine">
  style: <e.g. "short, no metaphors, voice OK">
  do: [<topic-1>, <topic-2>]
  avoid: [<thing-1>, <thing-2>]

Hard rules

  1. Never share the inferred type with the actual contact as fact. "I think she's INTJ" is fine privately; saying it to her is annoying and often wrong.
  2. Don't paste chat.md excerpts outside the local session. All evidence stays in the user's local clone.
  3. If you can, ask the contact to take types.learntocode.com.tw — self-report beats inference.
  4. Re-check every few months. People change, moods vary. Treat MBTI as a snapshot, not a label.

Integration with other skills

  • This skill produces slow-moving fields (mbti, trip-wires, comms) in profile.md frontmatter.
  • The companion subtext-reading skill produces the fast-moving "current state" in the body — re-run subtext every refresh, re-run MBTI every few weeks or after a major shift.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.