Source: LinkedIn post by Ruben Hassid ("Nobody has ever explained all of Claude to you", 2026-11-08, about 760 reactions and 125 comments 8 hours in). Text only, no infographic. Its one link is the author's skill library, skillsclau.de.
TL;DR
A consumer-level setup guide in six "levels", from the same author as R004, four days later. Four of the six are reasonable defaults; two are worth knowing even at staff level, and both check out against Anthropic's own pages: thumbs up/down sends the whole conversation to Anthropic for up to 5 years and it may be used for training, and on Team and Enterprise, an owner turning memory off for the organization permanently deletes every member's memory. The model-switching advice is right for the wrong reason (it is a cache miss, not a "re-read"). The post contradicts itself on skills: delete yours, then install his, from a 33,971-skill email-gated library. It also quietly reverses R004's "default to Auto" in favor of a fixed Opus 5.5 on high.
What the post says
Use the desktop app. It works with local folders without uploading files one by one, and it pulls you toward Claude Code.
Don't choose a model each time. Default to Opus 5.5 on high, and don't switch mid-chat, because "Claude re-reads the entire chat" and burns tokens.
Effort is Medium or High, nothing else. Medium for light work, High for most; Extra High mostly burns your usage limit.
Delete your old setup. Clear the custom-instructions box and old skills, since they were written for older models. Give one goal, and end prompts with "Before you start, restate my goal in your own words."
Ask for HTML, not a document. Add "Make it an HTML page" and "for [who will read it]". Claims there is then "no scope of making mistakes".
Protect what you plug in. Never click thumbs up or down, since feedback may go to training. On a Team plan, leave memory on once it's on: if an owner turns it off, every member's memories are deleted permanently.
Then a 20-minute plan: install the desktop app, no settings, connect one tool, use Claude Code more, delete installed skills and add new ones, and visit his library.
Claim check
| # | Claim | Verdict | Why |
|---|---|---|---|
| 1 | Desktop app beats the browser | Mostly right | Local folder access and Claude Code remove the upload step. The trade-off it skips: the agent can now read and write your disk, so permission prompts and folder scoping matter (R006's permissions by blast radius). |
| 2 | Don't switch models mid-chat; it "re-reads the chat" | Right advice, wrong mechanism | Every turn already sends the whole conversation; prompt caching is what makes that cheap. Caches are per model, so a switch is a cold cache: the full history is billed at full input price once, and again on any switch back. The cost scales with conversation length. A second, quieter cost: the new model inherits a transcript shaped by the old one's style. |
| 2b | Default to Opus 5.5 on high | Reasonable, unmeasured | Simple and stable, which also helps caching. It contradicts R004 (same author, 2026-11-04), which said to default to Auto. Neither post shows evidence. |
| 3 | Medium or High effort only | Plausible heuristic | Higher effort spends more thinking tokens; whether it pays depends on the task. Long agentic or hard-to-verify work can repay Extra High by avoiding retries; no data is offered either way. |
| 4 | Delete old instructions and skills | Same as R004, same gap | Pruning stale config after a model upgrade is sound, but wiping without a before/after eval also deletes project facts and rationale. Skills cost little until they trigger (only their descriptions sit in context), so the risk is a stale skill that triggers and overrides newer model behavior, not idle ones. |
| 4b | End with "restate my goal in your own words" | Good | A cheap alignment gate: misreads surface before work starts. It is the one-line version of EGM's comprehension test (R017). |
| 5 | Ask for HTML, "for [audience]" | Half right | Naming the reader is the real win. HTML gives rich layout but is harder to diff, edit and review than Markdown or a doc; it does nothing about factual errors. "No scope of making mistakes" is false. |
| 6a | Thumbs up/down can go to training | Confirmed | Anthropic's Privacy Center (consumer plans): feedback stores the entire related conversation, including content and style preferences, for up to 5 years, de-linked from your user ID, and it may be used to train models. It excludes raw connector content unless pasted into the chat. |
| 6b | Team owner turning memory off deletes everyone's memory | Confirmed | Claude Help Center: when an owner turns memory off for the organization, all existing memory entries for all users are deleted immediately and permanently. Owners can't view or edit members' memories. |
Takeaways for a staff engineer
Pin model and effort per conversation as cache hygiene. Choose up front; if a task needs a different model, start a fresh conversation with a short handoff rather than switching in place.
Treat the feedback button as a data-export path. For teams on consumer plans, write it into the AI usage policy: no thumbs on chats containing customer data, secrets or unreleased code.
Put the org-wide memory toggle under change control. It is a destructive, irreversible bulk delete behind a single setting; document it in the admin runbook and require a second person.
Version harness config instead of wiping it. Keep instructions and skills in a repo, re-audit them after model upgrades, and gate removals on a small eval (R005's quarterly re-audit).
Steal the restate-the-goal line. It costs one turn and catches the most expensive failure, solving the wrong problem.
Review checklist
Is the model and effort level chosen at the start of each long conversation, and not changed mid-thread?
Does the team's usage policy say when feedback ratings are allowed?
Who can change the organization's memory setting, and is that change reviewed?
Are instructions and skills versioned, with a dated reason for each rule?
Before deleting config, is there a before/after check on a few representative tasks?
Do prompts name the reader and ask the model to restate the goal?
Is local folder access limited to the folders a task needs?
Related entries
R004 (same author, 2026-11-04): the longer version of level 4. This post keeps the wipe-everything advice, drops "default to Auto" for Opus 5.5 on high, and adds the privacy points.
R017 (Elephant-Goldfish Model): "restate my goal" is a one-line form of its doc-alone comprehension test.
R005 (five-layer map): its quarterly re-audit with evals is the gate R004 and R018 lack.
R002: why a blanket wipe is risky; memory and rationale need maintenance, not deletion.
R006: the desktop app's local access is where permissions by blast radius start to matter.
Source notes
The post is paraphrased. Comments seen: a handful of short ones (praise, and one asking how a skill differs from an agent); none add technical substance. Privacy claims checked against Anthropic's Privacy Center article "Is my data used for model training?" (consumer plans; commercial plans have a separate article) and the Help Center article on chat search and memory. skillsclau.de was opened: a library of 33,971 skills, sorted by job, download requires an email sign-in; individual skills were not reviewed. Effort-level and caching statements are from general knowledge and are not re-measured.