Category: AI
Claude Code, AI tools, and practical AI workflows for developers and system administrators.
-
CLAUDE.md: Project Memory for Claude Code
Claude Code forgets everything between sessions unless you write it down. Where CLAUDE.md lives, what to put in it, /init, CLAUDE.local.md, .claude/rules/, and the auto memory Claude keeps for itself.
-
Claude Code Permissions: Allow and Deny Rules in settings.json
Stop answering the same permission prompt. The allow, ask and deny lists in settings.json, the rule syntax with the space before the *, PowerShell rules, protecting .env, the 3 files the rules live in, and the modes behind Shift+Tab.
-
Claude Code Hooks: Run a Check After Every Edit
A hook is a command Claude Code runs itself at a fixed point, every time, whatever the model decides. A PostToolUse hook on Windows that parses every .ps1 Claude writes and hands the errors straight back, plus the notification and block-the-file hooks.
-
Add an MCP Server to Claude Code
claude mcp add for a remote HTTP server and a local stdio one, the — separator, the 3 scopes and where each is stored, .mcp.json for the team, /mcp to check status and sign in, and the permission rule that stops the “use this tool?” prompts.
-
Resume a Claude Code Session With –continue and –resume
Closed the terminal mid-task? claude –continue picks up the last session in this folder, claude –resume opens the picker or a named session, /rename makes sessions findable, /branch tries a different approach, and /clear versus /compact.
-
Claude Code Subagents: Keep the Noise Out of Your Main Session
A subagent runs a task in its own context window and sends back only the summary. The built-in Explore and Plan agents, writing your own in .claude/agents/, limiting its tools and model, invoking one with @, and running a whole session as an agent.
-
Custom Slash Commands in Claude Code With Skills
Turn the prompt you paste every day into /name. A skill is a folder with a SKILL.md: where it lives, the frontmatter that controls who can invoke it, $ARGUMENTS, pulling live command output in with !`cmd`, and the PowerShell setting for Windows.
-
Plan Mode in Claude Code: Read First, Edit After
Plan mode lets Claude read the codebase and write a plan without touching a file until you approve it. Shift+Tab, /plan for one prompt, –permission-mode plan, the 3 approval options, editing the plan with Ctrl+G, and making it the default for a project.
-
Run Claude Code From a PowerShell Script With claude -p
claude -p runs one prompt and exits, which makes it a command you can put in a script. Piping a log in, getting JSON out and reading it with ConvertFrom-Json, allowing tools without prompts, –bare for a clean run, continuing a conversation, and the quoting that trips PowerShell.
Three guides collect the posts by the job you are doing rather than by date.
PowerShellFiles, CSVs, the firewall, scheduled tasks WSLInstalling it, distros, files across both worlds Windows AdministrationServers, domains, DNS, the tools worth havingThis site started as a SQL Server blog in 2017 and the archive is still here. The new SQL Server writing, and all the scripts, moved to a site of their own.
sqldba.blogScripts, error library, wait types, SSMS- CLAUDE.md: Project Memory for Claude Code
- Install Claude Code on Windows, Native and in WSL
- Claude Code Only Keeps 30 Days of Transcripts
- Claude Code Permissions: Allow and Deny Rules in settings.json
- Claude Code Hooks: Run a Check After Every Edit
- Compact a WSL or Docker Desktop vhdx to Get Disk Space Back
- AI (14)
- Linux & WSL (31)
- Microsoft SQL Server (51)
- Other (15)
- PowerShell (54)
- Windows Server Admin (40)