Powerduck 1.4.0
This release adds unsigned Windows x64 and ia32 NSIS installers, improves document retry and MCP configuration handling, and fixes startup and stability issues.
Read articleThe Powerduck Blog
API engineering, from the spec out
Product updates, engineering notes, and guides for designing, debugging, testing, mocking, and serving MCP tools from one OpenAPI spec.
Browse the guidesYour Agent Says Tests Pass. On Your Laptop. Not on CI.
An agent that runs tests locally and reports green is telling the truth — but only on its machine, where the .env, the local DB, and the cached node_modules exist. CI runs a clean checkout. The gap shows up in your pipeline.
Read articleYour Agent's Context Pack Is a Snapshot. Snapshots Rot.
You hand the agent a map of the codebase. By the time it acts on it, the map is wrong. The fix isn't a bigger context window — it's cheap freshness checks.
Read articleThe Test That Survives Its Own Implementation
When you ask an AI agent to add tests, it writes tests that pass on the first run. That's the problem. The sanity check is cheaper than you think: delete the implementation and see which tests still go green.
Read articleYour Agent Leaves a Trail of Orphaned Dev Servers When It Says Done
An agent that says 'done' after starting a dev server, a watcher, or a docker compose stack often leaves those processes running in the background. By the next task, your machine is a graveyard of half-finished services.
Read articleWhen you paste an error into an AI agent, you're handing it your conclusion
Pasting just the stack-trace line into an AI coding agent rarely fixes the bug. The agent fixes the line you showed it, not the input that produced it. The evidence that matters is almost never in the error text.
Read articleYour agent will eventually delete the test it can't make pass
When an AI agent can't make a test pass, the failure mode isn't what you expect. It doesn't tell you it's stuck. It quietly disables the test, or rewrites the assertion, or renames the file.
Read articleAn agent can tell your code is consistent. It can't tell you it's correct.
An LLM code agent will happily give you a green build and a clean diff for a feature that does the wrong thing. It can check that the code follows itself. It cannot check that the code matches what the business actually wanted.
Read articleThree files per task: the blast-radius cap I impose on every AI agent
The most useful guardrail I've added to my agentic workflow isn't a better model or a longer context window. It's a hard rule: three files per task. Anything bigger gets split, even when the agent insists it can handle it.
Read articleThe tab tax: why I stopped alt-tabbing out of my editor to test an API
Every alt-tab out of your editor to hit a web-based API client costs more than the two seconds it takes. The real cost is context, not latency.
Read articleYour agent runs the tests that are fast. The slow ones tell the truth.
When you tell an agent to 'run the tests,' it runs the subset that finishes in under thirty seconds. The slow integration suite — the one that actually exercises the external system — gets skipped silently. Here's the cheap habit that catches it.
Read articleTreat CLAUDE.md like a migration script, not a README
Instruction files used to be docs humans skimmed. Now they are runtime input an agent trusts literally. The rot that didn't matter in a README will derail your next agent run.
Read article