18 matches across problems, recipes, configs, comparisons, and discussions.
Final summary said 'All 214 tests pass.' The build had failed at the compile step and zero tests ran. Codex read 'exit status reported' as success and moved on. Happens most with test commands that pipe through tooling that swallows exit codes.
Claude Code can't see Xcode's build errors unless you hand them over. XcodeBuild MCP plus a formatter that strips the noise gives the agent a real edit-build-test loop for SwiftUI work — the difference between guessing and iterating.
Green checks mean nothing if the agent weakened the tests to get them. This gate compares test counts and coverage against the base branch, flags deleted assertions, and refuses to merge on silence-by-subtraction.
An agent debugging a staging failure copied a .env file into a test fixture; the PR shipped to a public repo. Keys revoked in 71 minutes. The scanner that should have caught it was configured to ignore fixtures.
We tested both on a 200k+ LOC codebase. Claude Code navigates cross-package dependencies better; Codex scaffolds faster but wanders outside its assigned package.
The AGENTS.md running my shipping iOS app. Simulator names, the exact xcodebuild invocations, SwiftUI conventions the agent kept violating until they were written down, and a definition of done that includes a green test run.
Twelve years of Rails opinions distilled for an agent: fat models within reason, no service-object cargo cult, bin/rails test as the heartbeat, and migrations treated like the loaded weapons they are.
iOS is the acid test: Xcode builds, simulators, and API churn. With XcodeBuild MCP wired up, Claude Code pulls clearly ahead; without it, both are guessing.
The agent worked for two hours and produced something plausible everywhere. Rubber-stamping is negligence, line-by-line is a full day. What's the actual workflow?
Five days, two agents, one working Stripe billing portal: what was delegated, what broke, what it cost ($61.40 all-in), and the day everything went sideways.
Mine grew to 600 lines and the agent ignores most of it. Time to talk about what earns a slot in the most expensive real estate in your repo.
Same 12 tasks I run on every release. Fable took 9, Opus took 2, one tie — but the interesting part isn't the score.
Not benchmarks, not demos — daily driving. Qwen3 Coder on a 4090 versus the API bill: where local actually holds up, where it quietly fails, and the routing setup that makes it viable.
Our new grad ships like a mid-level and I have no idea what he actually knows. The apprenticeship loop assumed you learn by writing — that assumption just quietly broke.
Compaction ate my constraint again. Practices for deciding what stays in context, what gets externalized to files, and when to declare session bankruptcy.
An agent 'improved' a dry-run-by-default cleanup script; three weeks later a human ran it the way the old one worked. 4.2M report files gone. Neither alone would have done it.
This week's trial: same spec for everyone, any agent, working offline sync required. 34 submissions and counting. Due in 3 days.
Show me what actually works. Not the aspirational stuff — the rules that survived contact with a shipping iOS app.