agentleFS
Sign inStart free
swayamdhawanswayamdhawan/x-agent/skills/x-audit/SKILL.md

x-audit

Skill0 starsChanged 38 days agoMITView on GitHub

Post-mortem on what the user has already published on X - which posts actually worked, why, and what to stop doing - from their own analytics export. Use when the user shares their X analytics CSV or past posts and asks "what's working", "why did this flop", "read my analytics", "audit my account", "what should I double down on", or wants to know whether links, threads or a posting time actually matter for them.

3 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions3 projects use a version of this.

    Word for word · 1

    1. Jakeschincariol/x-agent-skill★ 4100% same

    Edited · 0

    None found.

    Wrote their own version · 2

    1. lxyang20131208-star/x-algorithm-skills★ 1same name
    2. belzoralolbtw/x-agent-skill★ 0same 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 repository10 other files it gives its agents.

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:

    Add what the script cannot see, reading the top 5 and bottom 5 side by side:

    line 67

≈ “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.

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

x-audit

The only honest source of what works for an account is that account. Every rule in every X guide, including the ones in this pack, is a prior. The user's own last 30 posts are the evidence.

One tool lives in this folder and it runs:

python3 audit.py account_analytics_content.csv

Input

In order of preference:

  1. The analytics export. On desktop: x.com/i/account_analytics, then Export. It is a Premium feature. One row per post: impressions plus the engagement counts (likes, replies, reposts, bookmarks, shares, new follows, profile visits). X does not document the column names and they have changed before, so audit.py matches them loosely; the older analytics.twitter.com export parses too.
  2. A sheet by hand with post text, impressions, and whatever counts the user can see. Twenty rows is enough for a first pass.
  3. Screenshots of the posts with their numbers. Transcribe into a CSV.

Also read ~/.claude/x/log.md if it exists: it records which hook formula each post used.

What it measures, and why not impressions

Impressions are mostly a function of how many people already follow the user. audit.py ranks on things that are not:

metricwhat it tells you
Engagement ratewhether the post earned the reach it got
Replies per 1k impressionswhether it started a conversation. X's ranker weighs replies far above likes
Shares per 1k (reposts + shares)whether it travelled. Shares by DM or copied link are among the highest-weighted actions X publishes
Bookmarks per 1kwhether someone plans to use it
Follows per 1kwhether it made anyone want more of the user

Posts under a noise floor of impressions are left out of the ranking, and the user's replies on other people's posts (rows starting with @) are counted and set aside so a good week of replying does not drown the posts.

Then find the pattern

audit.py groups the posts by hook formula (classified with ../x-post/hooks.json), by length, and by format (text only vs with a link or media), and reports every group whose engagement rate is 1.4x above or below everything else, with its n and a confidence word. It checks day of week last, and only mentions it when nothing else separates. It is almost never the cause, and it is where people want it to be.

Add what the script cannot see, reading the top 5 and bottom 5 side by side:

  • Topic. Which of the user's themes are in the top five?
  • Whether the user replied in the first hour.
  • Whether the post was the user looking good or the user looking bad.

The link question, answered with their own data. X says it stopped down-ranking link posts over a year ago. An independent analysis in January 2026 still measured fewer views on them. If the format finding shows link posts well below the rest with n of 5 or more, that is the answer for this account, whatever anyone says. If it shows nothing, links are fine here.

State every finding as a claim with the evidence attached, and say how confident it is. With 30 posts a pattern can show. With 8, say that instead of inventing one.

Output

AUDIT  ·  38 posts  ·  Jul 01 - Aug 28  ·  15 replies set aside

TOP BY ENGAGEMENT RATE
    5.8%    8.6 rep/1k    8.6 shr/1k     5,705 imp  "$4,000 is what one bad hire cost me."
    ...

WHAT THE DATA SAYS
  1. format = text only: 2.7% vs 0.8% for everything else (3.3x, above). n=26, solid.
  2. hook formula = Cost Confession: 3.9% vs 1.9% (2.0x, above). n=4, weak.
  Day of week shows nothing. Stop optimising it.

STOP:     link-first posts. Put the link in a reply for the next ten and re-run.
DO MORE:  the posts where you paid a cost, with the number in line 1.

Then hand the conclusions to /x-plan, so next week is built on the user's own evidence instead of on defaults.

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.

Check your setup, free