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

self-mirror

Skill0 starsChanged 6 months agoApache-2.0View on GitHub

Quantify the user's own WeChat chat habits across all tracked contacts — echo-reply count, opening-line repertoire, length distribution, time-of-day pattern, connector word frequency, emoji/sticker usage. Activate when the user asks "how am I actually talking on WeChat?", "do I sound like a bot?", "show me my豆包式聊天 frequency", "how do I open conversations?", or before a hard relationship conversation where the user wants to see their own pattern. The single most actionable target is echo-reply detection ("X 看起来挺 X 啊" pattern) — the canonical "你跟豆包聊天" failure mode. Wraps tools/self-mirror.ps1 which auto-detects the user's own sender label and produces a 7-section Markdown report.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 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.

  • 0

    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 quantitative analysis of the user's own outgoing messages — NOT the contacts' messages.

    line 8

≈ “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

Self-Mirror Skill

When this skill is loaded, you are running a quantitative analysis of the user's own outgoing messages — NOT the contacts' messages. The point is to surface uncomfortable truths about how the user actually communicates.

Tool: tools/self-mirror.ps1 does the heavy lifting. You (the agent) interpret the report.

How to invoke

.\tools\self-mirror.ps1                                # generate SELF-MIRROR.md (gitignored)
.\tools\self-mirror.ps1 -Person alice                  # narrow to one contact
.\tools\self-mirror.ps1 -Out reports/SELF-MIRROR.md    # custom output path

The tool reads people/*/chat.md, auto-detects the user's sender label, and writes SELF-MIRROR.md at repo root (gitignored).

What the report contains (7 sections)

  1. Echo-reply offenses — sentences matching "<X> 看起来挺 <X>" patterns where user restates the contact's prior message + softener. Counts + top 10 examples with date + recipient.
  2. Top 10 opening lines — most frequent thread-starters (new thread = >2h gap from prior in that chat).
  3. Length distribution — user-outgoing message char-length histogram, per-recipient.
  4. Question vs statement ratio — % ending in ? / ?, per-recipient.
  5. Time-of-day pattern — hour-of-day histogram, with do-not-disturb-window violations flagged (per recipient's comms.frequency).
  6. Connector/softener word frequency — 嘛 / 吧 / 啊 / 呢 / 哦 / 嗯, per-recipient. Flagged if >5% of words for a recipient.
  7. Emoji/sticker proxy count — emoji chars + bracketed sticker descriptions, per-recipient.

How to read the output

The killer insight is usually section 1 (echo-reply) — if the count is high, that's the user's豆包式 problem made concrete with dates and recipients.

Other patterns worth surfacing in your interpretation:

  • Length asymmetry: user writes 300+ chars to a contact who replies in 10 chars → user is over-explaining
  • Time-of-day violations: user messages a 22-点-不发 contact after 22:00 → wrong-mode-of-respect
  • Opening line repertoire of 1-2: user has no opener variety → predictable, easy to ignore

Hard rules

  1. SELF-MIRROR.md is gitignored — it contains real names + real message excerpts. Never commit.
  2. Show the data, don't moralize. Counts + examples + dates. The user decides what to change.
  3. Don't suggest specific replacement phrasings. The user's voice should stay the user's voice. If they ask for replacement, defer to subtext-reading per-contact context.
  4. Re-run after a behavioral correction attempt (e.g., 2 weeks after the user says "I'll stop echo-replying"). Compare before/after counts.

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.