I Built CI for My AI Agent (It Catches What You Miss)
A prompt change broke my agent silently. Behavioral CI caught 4 regressions before they shipped.
May 16, 2026 · 8 min
Archive
36 posts · 2025–2026
Latest — recent dispatches
How I Debug AI Agents Like Code (Not Guesswork)
10 trace CLI commands that turn re-run-and-guess debugging into actual inspection.
May 09, 2026 · 6 min
I Cut My AI Agent's Token Costs 21% Without Changing the Model
My agent burned tokens on outputs it never reread. One context change cut costs 21%.
May 02, 2026 · 8 min
How I Added Observability to My AI Agent (Without a Hosted Dashboard)
Failures I couldn't reproduce led to a trace-first layer. No dashboard, no infra, just traces.
April 25, 2026 · 9 min
Hybrid Search vs Query Routing for AI Agents: Benchmark Results
The router missed exactly when it mattered. Benchmarked against hybrid RRF on real PDFs.
April 18, 2026 · 11 min
How to Ship an MCP Server to Production (Authentication, Safety, and What Breaks)
MCP tutorials stop at tool integration. The real work is auth, guardrails, and untrusted data.
April 11, 2026 · 7 min
Semantic vs Keyword Search for AI Agents: When to Use Each
Semantic search finds synonyms. Keyword search finds invoice numbers. Benchmarked in pdf-mcp.
April 07, 2026 · 8 min
How Claude Code Actually Reads PDFs: Lessons from Building an MCP Server
Tool boundaries shape agent behavior more than prompts. What I learned shipping pdf-mcp.
April 04, 2026 · 10 min
My AI Agent Passed Every Check. 67% of It Was Wrong.
Every check passed. 67% of the facts were wrong. Three layers to catch confident hallucinations.
March 28, 2026 · 10 min
No articles found for this filter.