Sitemap
Every surface, with counts where the library grows.
Learn
Explore
The School
Articles (13)
- Build a Working Agent with sovereign-agent
- Build Your First ZeoCore Tool and Check Its Failure Boundaries
- Govern a Multi-Agent Project With Zero-Employee
- Why AI Agents Don't Dissolve the Firm
- Your Claude Code Permissions Are Your Org Chart
- Clean a Messy Customer List With Claude Code — and Prove Nothing Was Lost
- Price the Agent Work You Can Verify
- What Claude Code Actually Costs — and Which Dials Move It
- What Claude Code Actually Changes — and What It Doesn't
- When Can You Actually Walk Away?
- Your Chat History Is Not a System of Record
- The Content Kit: Every Block, and When to Use It
- What Is a Zero-Employee Organization?
Thoughts (62)
- Give a Reasoning Model a Clear Destination
- Record the Model Alongside the Prompt
- Write the Evaluation Before Tuning the Prompt
- A Valid JSON Object Can Still Be a Wrong Answer
- Test a Model on the Exceptions You Actually See
- Keep a Codex Handoff Smaller Than the Session
- Ask Codex to Locate the Source of a Generated File
- Turn a Repeated Codex Task Into a Tested Skill
- Check Codex's Exit Status Before Using Its Output
- Use Codex on a Small, Known Migration First
- Give a Codex Skill One Example It Must Reject
- Save the Codex Artifact Before Clearing the Chat
- Retire a Temporary Codex Instruction Override
- Separate Codex's File Access From Its Network Access
- Write Review Rules Around Business Invariants
- Keep a Codex Review Finding Reproducible
- Make the Codex Working Directory Explicit
- Give an Automated Codex Report a Schema
- Keep Codex Progress Out of Your Parsed Result
- Check a Codex Worktree Before Trusting Its Result
- Give a Codex Skill a Clear Trigger
- Make Codex Explain Which Instructions It Loaded
- Review the Change That Was Never Requested
- Teach Claude the Difference Between Blank and Zero
- Measure the Cost of an Accepted Result
- Keep a Local Setting Out of the Team Contract
- Make a Hook Prove Its Failure Behavior
- Treat an MCP Tool List as an Access Review
- Check What a Permission Rule Actually Allows
- Test the Build Your Customer Will Receive
- Trace One Failed Request Through the Logs
- Preview Every Rename Before Moving Files
- Give Claude a Broken Example Before a Fix
- Make Claude Preserve the Meaning of a CSV
- Give the Agent an Inspectable Check
- After Repeated Corrections, Rebuild the Task Context
- Test Whether a Shorter CLAUDE.md Helps
- Delegate Discovery With an Evidence Handoff
- Know What Rewind Can Restore
- Use an Interview to Expose Gaps in the Brief
- Put Repeatable Checks in Command Hooks
- Make Completion Checks Explicit, and Test Their Limits
- Give the Reviewer a Fresh View of the Change
- Triage Review Findings Against the Requirement
- Budget the Coordination Before Adding a Team
- Bound Each Run Before Scaling a Migration
- Parameterize the Brief, Verify the Execution
- Choose the Subagent Model Against the Task
- Keep the Handoff Outside the Summary
- Separate the Checkouts, Then Integrate the Work
- Test What a Different Model Adds to Review
- Review the Render Against the Requirement
- Give the Agent a Precise Source to Inspect
- Separate Permission Decisions From Execution Boundaries
- Turn Repeated Setup Advice Into Maintained Instructions
- Plan When the Approach or Consequences Are Unclear
- Keep a Side Question Out of the Main Thread
- Separate Shared Instructions From Local Context
- Keep a Session Focused, and Check Its Working Tree
- Pass the Evidence Without the Clipboard Handoff
- Choose an SDK When You Need to Own the Runtime
- Check What the Review Command Actually Runs
Courses (11)
- Agentic Coding with Cursor
- Claude Code for Business
- Why Agents Fail: Harness Engineering for AI Projects
- Tools, Memory, and Multi-Agent Systems
- Two Architectures: The 2x2 for Agent Design
- Voice Agents, Not Chatbots
- The 12-Week Transformation Tracker, Run on Claude Code
- ZEO ITAM · Autumn 2026
- NVIDIA NIM in Production: From Hosted API to Self-Hosted Inference
- LangGraph Agent Orchestration
- Nebius Token Factory in Production: Running Open Models Honestly
