x-link-best-effort
Skill0 starsChanged 6 months agoMITView on GitHub
Recover article, thread, and post drafting from x.com links even when direct X page fetching is blocked. Use when an X status/article link is the source and the user wants analysis, an article, a post, a thread, or a content draft.
In this repository15 other files it gives its agents.
- approved-backlog-watchdogSkill
- context-engineeringSkill
- elon-operatorSkill
- goal-modeSkill
- image-readingSkill
- local-article-toolSkill
- mac-ttsSkill
- personal-knowledge-wikiSkill
- repurpose-funnelSkill
- skill-builderSkill
- swipefile-compilerSkill
- thumbnail-creatorSkill
- video-to-skillSkill
- visual-first-explainerSkill
- youtube-learnSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
21 rules. About 0 are common, 11 are distinctive, and 10 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈11
Distinctive · estimate
This author's own choices. Worth a look.
10
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 8
Use this skill when the task starts from an x.com or twitter.com link and the goal is to learn from it, analyze it, or turn it into a new …
- Distinctive: line 12
Do not stop at the X block screen.
- Distinctive: line 14
If direct X page fetching is blocked, continue with the best evidence already available and produce the requested output immediately.
- Very specific: line 18
Use x_tweet_fetch first for status or article links.
- Very specific: line 19
If the payload contains article_title, article_preview_text, or article_plain_text, treat that as the primary source article.
- Very specific: line 20
Only use web_fetch for non-X URLs discovered from the tweet payload.
- Very specific: line 21
Do not use direct web_fetch on x.com as the first attempt.
- Very specific: line 22
Use direct browser fallback on x.com only if Charlie explicitly asks for deeper retrieval.
- Very specific: line 26
When the visible fetch result is just the X error shell such as Something went wrong or the privacy-extension warning:
- Distinctive: line 28
Do not ask Charlie to choose between pasting text and letting you infer.
- Very specific: line 29
Do not ask Which do you prefer?
- Distinctive: line 30
Do not present an A/B menu.
- Distinctive: line 31
Do not stop at the extraction failure.
- Very specific: line 36
Use whatever evidence you already have: tweet text, article fields from x_tweet_fetch, link title, snippet, author, visible metadata, …
- Distinctive: line 42
If the user asks for an article, post, thread, or draft:
- Distinctive: line 44
Write the deliverable immediately.
- Distinctive: line 45
Treat the source as learning material, not a reason to pause the workflow.
- Very specific: line 46
If the user says learn from it, write a new original piece instead of a summary unless they explicitly ask for a summary.
- Distinctive: line 47
Do not copy source wording or source structure unless the user explicitly asks for preservation.
- Distinctive: line 48
When evidence is partial, write the piece with honest confidence: keep claims tight, avoid fabricated specifics, and make any inference …
- Distinctive: line 67
no option menu unless Charlie explicitly asks for alternatives
≈ “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
X Link Best Effort
Use this skill when the task starts from an x.com or twitter.com link and the goal is to learn from it, analyze it, or turn it into a new draft.
Core rule
Do not stop at the X block screen.
If direct X page fetching is blocked, continue with the best evidence already available and produce the requested output immediately.
Retrieval order
- Use
x_tweet_fetchfirst for status or article links. - If the payload contains
article_title,article_preview_text, orarticle_plain_text, treat that as the primary source article. - Only use
web_fetchfor non-X URLs discovered from the tweet payload. - Do not use direct
web_fetchonx.comas the first attempt. - Use direct browser fallback on
x.comonly if Charlie explicitly asks for deeper retrieval.
If X blocks read-only fetching
When the visible fetch result is just the X error shell such as Something went wrong or the privacy-extension warning:
- Do not ask Charlie to choose between pasting text and letting you infer.
- Do not ask
Which do you prefer? - Do not present an A/B menu.
- Do not stop at the extraction failure.
Instead:
- State briefly that direct X page fetch was blocked.
- Use whatever evidence you already have: tweet text, article fields from
x_tweet_fetch, link title, snippet, author, visible metadata, surrounding user instruction. - Clearly separate
Fetched,Inferred, andUnknownonly when that distinction helps. - Then complete the requested deliverable in the same reply.
Drafting contract
If the user asks for an article, post, thread, or draft:
- Write the deliverable immediately.
- Treat the source as learning material, not a reason to pause the workflow.
- If the user says
learn from it, write a new original piece instead of a summary unless they explicitly ask for a summary. - Do not copy source wording or source structure unless the user explicitly asks for preservation.
- When evidence is partial, write the piece with honest confidence: keep claims tight, avoid fabricated specifics, and make any inference conservative.
Preferred phrasing
Good:
Direct X fetch was blocked, but I could still extract enough signal to build the draft.I used the available tweet metadata and visible article clues, then wrote the article below.
Bad:
I still can’t fetch X reliably from my side.Two options so I can move immediately:Paste the tweet/thread text here.Which do you prefer?
Output policy
Default to momentum:
- brief note on evidence quality
- requested analysis or draft
- no option menu unless Charlie explicitly asks for alternatives
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.

