Agentic Workflows — Mastery7 / 8
The Daily-Driver Setup — Settings, Permissions, Status Line
The difference between fighting the agent and flowing with it is twenty minutes of configuration you do once. Here is the setup.
Most friction with an agent is unconfigured friction: a permission prompt on the hundredth npm test, or a destructive command you wish it had refused. Twenty minutes in settings.json fixes both.
The settings hierarchy
Settings merge, most-specific winning: enterprise → user (~/.claude/settings.json) → project (.claude/settings.json, committed) → local (.claude/settings.local.json, git-ignored). Put team conventions in project; keep secrets in local or real env vars.
A permission baseline worth committing
{
"permissions": {
"allow": ["Read", "Edit", "Bash(npm run test:*)", "Bash(git status:*)"],
"deny": ["Bash(rm -rf:*)", "Read(./.env)", "WebFetch"]
}
}
allowyour safe, high-frequency commands so the session flows.denythe genuinely destructive —denyalways wins overallow.- Leave everything else to prompt.
Permission modes
default— prompt on first use of each sensitive tool.acceptEdits— auto-accept file edits, still prompt for other risky tools.plan— read-only: explore and plan, change nothing.bypassPermissions— run everything, no prompts. Powerful and dangerous — sandboxed, fully-trusted contexts only.
A status line you read
Configure statusLine to show model, branch, and token cost. When you can see what the agent is using at a glance, you catch the wrong model or a runaway context before it costs you.
Setup done. The finale: chaining MCP, hooks, commands, and subagents into a pipeline that takes a ticket to a reviewed branch.
Series — Agentic Workflows — Mastery
- Part 01The Agentic Workflow MindsetMost developers treat AI like a smarter autocomplete. The ones who pull ahead treat it like a teammate that acts. Here is the mental model.
- Part 02MCP Servers 101 — Give Your Agent Real ToolsThe Model Context Protocol is how your agent stops guessing and starts querying your database, your issues, your browser. Here is the mental model and the first connection.
- Part 03Build Your First Custom MCP ServerOff-the-shelf servers cover GitHub and Postgres. The high-leverage one is the server only you can write — the bridge to your own system.
- Part 04Hooks — Make the Agent Obey Your RulesA prompt asks the model to remember. A hook makes it happen — deterministically, every time, outside the model's control.
- Part 05Custom Slash Commands as Team WorkflowsA custom slash command is a reusable prompt you commit to the repo — so the whole team runs the same high-quality instruction instead of re-typing it.
- Part 06Subagents — Delegating Work That ScalesOne giant context gets slow and vague. Subagents let the main agent hand focused work to specialists with their own context and tools — and run them in parallel.
- Part 07The Daily-Driver Setup — Settings, Permissions, Status Line — you are hereThe difference between fighting the agent and flowing with it is twenty minutes of configuration you do once. Here is the setup.
- Part 08Multi-Tool Pipelines — Ticket to Reviewed BranchThe payoff: chain MCP, hooks, commands and subagents into one flow that takes a ticket to a reviewed branch — with you in the loop only where it counts.