AI
Should AI skills follow the org chart or the process?
AI skills need both. A skill belongs to a role, which looks after it and may change it, and a step in a written process names it, which settles when it is used. Anthropic groups skills into role bundles. Filed by role alone, people hold tools with no order of work, and filed by process alone, nobody looks after them.
Is a consumer Claude plan safe for client data, and is it cheaper than Team?
Claude Max and Ultra run under the Consumer Terms, which carry no data processing addendum, so I keep client personal data on Team or Enterprise. On price, Max costs the same per five-hour session as a Team seat billed monthly. What Ultra buys is a higher ceiling for one heavy Claude Code user.
Agent-ready website: what I built on two sites and what has evidence
An agent-ready website, as I built it on tallyfy.com and this site, means llms.txt, schema, WebMCP tools and discovery files. Ahrefs found 97% of llms.txt files got zero requests in May 2026. Some of those parts have evidence behind them, and I took out several that did not.
Do you need a director of ontology for AI agents?
A director of ontology builds the shared vocabulary AI agents reason over. Capital One, Novartis and Lockheed Martin are hiring for the work, yet UK ads citing ontologies were 0.035% of permanent job ads in the six months to October 2026. Most companies should start with typed files instead.
The LLM writes, the model decides
Pair a writing LLM like Claude with a cheap decision model that only returns numbers, such as TypeSafe Jev. Three ways to split the work, why a probability Claude writes down is a self-report and not a measurement, and why a reason written after the fact for a bare number is worse than no reason.
Claude inference hooks cannot see a screenshot
Anthropic now lets an Enterprise organisation put its own server in front of every prompt and return allow or deny before the model runs. It is the strongest inline control Claude has shipped. It also never receives raw image bytes, so a screenshot of the document you are trying to stop walks straight through, and the failure mode when your server goes down is a setting somebody has to choose.
Claude US-only inference costs 1.1x and that is the cheapest part
The Enterprise console offers one toggle that keeps all model inference in US regions for a 10 per cent surcharge. Underneath it are two independent settings, one of which is a one-way door you set when you create a workspace and can never change. The surcharge is the part everybody reads and the least expensive of the three costs.
The Claude Enterprise admin console has 25 sections and no map
Claude Enterprise puts 25 settings sections behind one nav, grouped into four blocks that do not match how anybody actually governs AI. Here is the whole tree, what each block controls, and the six settings worth opening on your first day as an Owner rather than the twenty-five you will otherwise scroll past.
macOS TCC broke brew, mise and Claude Code at the same time
Three tools failed within a minute of each other with three unrelated errors, and the folder they blamed had ordinary permissions the whole time. macOS TCC had revoked it. Two tccutil resets fixed that with no restart, because Terminal and Claude Code hold separate TCC identities and each one needs its own grant.
Claude for Excel cannot save a file, and the error will not tell you that
Ask the Claude add-in in Excel to roll a workbook forward into next month and you get a permissions error, which reads like a misconfiguration and is not one. The add-in has no file management at all, and that limit is missing from the unsupported list. Here is what it does do well, why it beat Claude Code on the same spreadsheet, and the settings field that fixes the complaint everybody has.
Why your Claude Microsoft 365 connector is still read only
Anthropic shipped write tools for the Microsoft 365 connector on July 7, 2026, and left them switched off for every organization that had connected before that date. Nothing in the product says so. Here is how to check in ten seconds, the two admin gates in two different consoles, the scope that deserves a second look, and the write path already running in your building that will derail the diagnosis.
How to run many Claude Code sessions without duplicate work
Claude Code sessions can now message each other over a Unix socket on your own machine. I had 23 sessions running and only 14 of them could see each other. Here is what the message channel fixes, what it cannot fix, and why the guard I trusted for months was never a guard at all.
One line in your org-wide AI instructions cuts output by 24 percent
Adding an organization-wide instruction block does add tokens to every prompt, and that objection is correct. A controlled 180-call test on Claude Sonnet 5 shows one brevity line cutting output 23.9 percent and returning 14.2x its own cost, while the block carrying it fails to break even uncached.
Why a Claude share link will not open, and what actually works
A shared Claude conversation returns HTTP 200 whether or not you can read it, and the page carries six characters of visible text. Team and Enterprise chats are organization-only by design, attached files never travel with the snapshot, and the one programmatic route is a Compliance API that has no idea what a share link is.
Two of your agents never read your CLAUDE.md
Claude Code loads your CLAUDE.md into every subagent except two. Explore and Plan skip it by design, and no setting changes that. So the agents you fan out most widely are the ones that never see your rules. ETH Zurich measured what the file costs on the occasions it does load.
A merge queue is theatre without a test oracle
Merge queues, merge trains, and speculative merging all rest on one assumption nobody states: that a per-branch green tick means the code works. Our own CI runs no unit tests at all, and master went red twice in two days from pull requests that were each green on their own. Here is what to build before you buy the queue.
What a git worktree does not isolate
A git worktree isolates your working directory and your index. It does not isolate the stash stack, which is shared by every checkout in the repository. Here is what I measured on git 2.50.1, why MERGE_AUTOSTASH escapes the problem through an accident of spelling, and how one defensive stash can make a worktree look safe to delete.
An MCP server is unreviewed code with your file system in scope
Treat every MCP server as untrusted code that runs with the access your agent has, because that is what it is. Anthropic docs say the directory lists connectors but does not security-audit them. A registry of approved servers with nothing enforcing it is a memo. The control that binds is a managed allowlist matched by URL or command, never by name.
Claude is allowed in regulated finance, but it has no EU data residency
Two objections kill most regulated-finance AI conversations before they start. The first, that Anthropic does not permit Claude for regulated work, is false: Claude for Financial Services exists, banks run it, and the usage policy names finance high-risk, not forbidden. The second is real and almost nobody states it plainly: first-party Claude Enterprise has no EU data residency at all. There is no "eu" inference region and workspace storage is US-only. If you are FCA-regulated, that is the fact to design around, and the only EU route runs through a hyperscaler.
Your Claude Code deny rules are not a security boundary
Before you hand Claude Code to hundreds of people you add deny rules for .env and credentials and feel locked down. You are not. Those rules govern Claude own tools, not a Python one-liner that opens the same file, and the control that actually holds, the OS sandbox, reads your whole machine by default and fails open when it cannot start. The baseline worth setting is real. Its dangerous gaps are the defaults you never changed.
Your locked-down Claude sandbox is a holding pattern, not a destination
Giving everyone Claude inside an isolated VM, no sensitive data allowed, feels like the safe way to start. It is a fine way to start. The trouble is what happens when you leave people there: the leak it was built to stop walks out by copy-paste anyway, the friction recruits the shadow AI you were trying to prevent, and the value never compounds because nothing in an ephemeral box survives the session. A sandbox is a scaffold. Scaffolds come down.
Blocking the personal Claude account is an identity problem, not a network one
Your CISO trusts the control posture Microsoft gives Copilot. To get Claude to the same bar, do not reach for tenant restrictions: that header only fires on your network, so it is theater the moment a laptop goes off-VPN. The control that holds lives at identity. Enforce SSO, then claim your domain, and know that the claim is a one-way door.
Claude Code loop is for the work you watch
Claude Code /loop reruns a prompt on an interval inside your session. It is perfect for babysitting a deploy or a test run, and wrong for anything that has to keep going while you are away. It also exposes a durable flag that, on version 2.1.185, quietly writes nothing to disk. Here is how to use it well.
How to schedule Claude Code on your own machine
You want a Claude job to run every few hours on your Mac, not in the cloud. A cloud routine cannot do it, because it never touches your machine. Here are the local options that can, why launchd beats cron for this, and a working LaunchAgent that pulls every one of my repos on a schedule.
Accessibility overlays do not work, and AI auditing is the opposite
An accessibility overlay is one line of JavaScript that promises ADA compliance while you do nothing. The FTC fined accessiBe a million dollars over that promise. Here is why a widget cannot fix a problem that lives in your code, and how real AI auditing does the reverse by finding the broken line so a person can change it.
Can AI actually do accessibility testing? I ran it on my own product
Automated accessibility tools catch maybe a third of WCAG problems. I pointed Claude Code at Tallyfy, my own product, and let it run a real WCAG 2.2 audit with a live screen reader across four codebases. It found bugs that axe-core cannot see, and it showed clearly where the work still needs a person.
Claude Code behind a TLS-inspecting proxy: configure the tool, not the proxy
Locked-down shops reach for a proxy exception to make Claude Code connect. Wrong move, and it fails anyway. Claude Code does not pin certificates, so it works through full TLS inspection once you teach it to trust your corporate root CA. The fix is a couple of environment variables and an egress allowlist, not a hole in the proxy.
How to run a long autonomous Claude Code job without it drifting
The hard part of a big AI job is not the work. It is making the agent run for many sessions without drifting or claiming it is done when it is not. I used an accessibility audit across four codebases as the test. The setup that kept Claude Code on track was a git ledger, atomic parallel claims, and two verification passes.
What a VPAT costs, and why the report is the cheap part
A VPAT is the report that states how accessible your product is, measured against WCAG. People ask what it costs and price the document, but the document is the cheap part. The real cost is re-auditing every release, and that is the number an AI agent actually moves. Here is the ADA, WCAG, Section 508 and EN 301 549 stack underneath it.
What axe-core misses, and how AI caught it with a real screen reader
Axe-core catches about a quarter of WCAG failures and skips anything that needs judgment. Here are the thirteen criteria a scanner cannot decide, how an AI agent drives a real VoiceOver session to cover them, and the save button that passed every automated check and was silent to a blind user.