A PreToolUse hook blocks git reset --hard and force pushes
A rule in CLAUDE.md is a request. A hook that exits 2 before the command runs is a guarantee.
Claude CodeGitverified 2026-11-21
Reviews from people who ran it with a coding agent: the exact setup, the result, and when someone last checked it.
A rule in CLAUDE.md is a request. A hook that exits 2 before the command runs is a guarantee.
Claude CodeGitverified 2026-11-21
Reuse, the standard library and the platform come before new code. Its benchmark shows about half the code at the same correctness, with clear caveats.
Claude CodeCodexAnyverified 2026-11-21
The agent walks the decision tree of your plan in rounds, gives a recommended answer for each question, and builds nothing until you both agree.
Claude CodeAnyverified 2026-11-21
One failing test, then the code to pass it, then the next. Tests written in bulk check imagined behavior, and an agent can fake its way past them.
Claude CodeAnyverified 2026-11-21
A "lazy senior developer" skill. The agent reuses what exists, skips features nobody asked for, and says what it skipped.
Claude Codeclaude-opus-5-5TypeScriptverified 2026-11-18
Two fallow commands in the pre-commit hook catch layer violations and dead code the agent leaves behind, before I review anything.
Claude Codeclaude-opus-5-5TypeScriptRemix 3verified 2026-11-18
An agent stops when the work looks done. Give it a command that says pass or fail, and wire that command into a hook so "done" means green.
Claude CodeTypeScriptverified 2026-11-21
A planning interview about a whole product fills the context window before the plan is done, and answers get worse as it fills.
Claude CodeAnyverified 2026-11-21
Shouting at the model adds tokens, not reliability. A rule with no exceptions belongs in a hook or a deny rule.
Claude CodeAnyverified 2026-11-21
Every line in the instruction file costs attention on every task. Keep only what is both undiscoverable and always relevant, and point to the rest.
Claude CodeCodexCursorAnyverified 2026-11-21
A bash loop feeds the same prompt again and again. Each pass picks one open task from the PRD, checks, commits and logs progress, until the list is done.
Claude CodeAnyverified 2026-11-21
Split the work into three sessions with a written handoff each. A human reads 200 lines of plan, not 2000 lines of diff.
Claude CodeAnyverified 2026-11-21
A second agent with a fresh context and no write tools checks the change against SPEC.md, and reports only gaps that affect correctness.
Claude CodeAnyverified 2026-11-21
When the agent gets something wrong, fix the harness, not just the output, so the same mistake cannot happen in the next session.
Claude CodeCodexAnyverified 2026-11-21
Two sessions in separate worktrees each rewrote the same scoring rule in a different way; the second push was rejected and the rebase conflicted, and only the agent stopping to ask kept one version from silently overwriting the other.
Claude Codeclaude-opus-5-5TypeScriptverified 2026-11-21
Instead of prose rules, AGENTS.md opens with a two-column table. The agent finds the row for its task and reads only the file that row points to.
Claude Codeclaude-opus-5-5TypeScriptRemix 3verified 2026-11-20
Every parallel agent gets its own git worktree and its own Neon branch. Agents never share a checkout or a database, and none of them sees production.
Claude Codeclaude-opus-5-5PostgresNeonGitverified 2026-11-20
Remix 3 is not React, but the model's memory says it is. One line in AGENTS.md sends the agent to node_modules/remix/INDEX.md first, and the React-shaped code stopped.
Claude Codeclaude-opus-5-5TypeScriptRemix 3verified 2026-11-20
Every task ends with the same four steps and a reply that names what was skipped. The agent stopped saying "done" on a red build.
Claude Codeclaude-opus-5-5TypeScriptVite+verified 2026-11-20
One hard rule in AGENTS.md stops narrating comments and summary files that go stale the day they are written.
Claude Codeclaude-opus-5-5TypeScriptverified 2026-11-18