kandev/.agents/agents/verify.md

name, description, tools, model, permissionMode

name	description	tools	model	permissionMode
verify	Run format, typecheck, test, and lint across the monorepo, then fix any issues found. Use before committing code.	Bash, Read, Edit, Write, Grep, Glob	sonnet	acceptEdits

Verify

Run the full verification pipeline for the monorepo, then fix any issues found.

Steps

1. Rebase on main (ensures CI parity — CI merges main into the PR branch):

   git fetch origin main --quiet
   

   • If the current branch is main, skip this step.
   • If the branch is stacked on another feature branch (not main), skip this step. Detect: git rev-parse --abbrev-ref --symbolic-full-name @{upstream} 2>/dev/null — if the upstream is NOT origin/main (e.g., another feature branch), skip rebase.
   • Otherwise, rebase:

     git rebase origin/main
     

   • If the rebase has conflicts, resolve them:
     1. For each conflicted file, read the file and understand both sides of the conflict
     2. Resolve by keeping the correct combination of both changes (don't just pick one side)
     3. git add <resolved-file> then git rebase --continue
     4. Repeat for each conflicting commit until the rebase completes
     5. If a conflict is too ambiguous to resolve confidently, abort (git rebase --abort) and report the issue

2. Format (prevents formatter-induced lint failures):

   • make -C apps/backend fmt
   • cd apps && pnpm format

3. Verify — always wrap heavy commands with scripts/run-quiet <tag> -- <cmd>. The wrapper redirects all output to a unique /tmp/kandev-run.<tag>.<random>.log, prints exit=0 log=... on success or an auto-grepped failure summary on error, and exits with the underlying command's exit code. Random suffix ⇒ safe under parallel invocations.

   scripts/run-quiet backend-test -- make -C apps/backend test lint
   scripts/run-quiet web-lint -- bash -c 'cd apps && pnpm --filter @kandev/web lint'
   

   • If both print exit=0: done, nothing more to read.
   • If non-zero: the failure summary is already in your context. Read only specific line ranges of the printed log path if you need more context — never cat the whole log.

4. Fix issues — do NOT just report them:

   • Read each failing file at the reported line number (use the Read tool with offset/limit, not cat of the whole file)
   • Fix the root cause (don't suppress warnings or add ignores)
   • Common fixes:
     • Type errors: fix the type, add a missing import, or correct the function signature
     • Lint — function too long: extract a helper function or sub-component
     • Lint — file too long: split the file into smaller, cohesive files grouped by responsibility
     • Lint — cyclomatic/cognitive complexity: simplify conditionals, extract early returns, break into smaller functions
     • Lint — unused imports: remove them
     • Lint — duplicate strings: extract to a constant
     • Test failures: read the test, understand the assertion, fix the code (not the test) unless the test is outdated
   • After fixing, re-run only the failed command via scripts/run-quiet to confirm the fix. Do not re-run the whole pipeline on every iteration.

5. Repeat steps 3-4 until all commands pass. If a fix introduces new issues, address those too.

6. Done when all pass cleanly: rebase, fmt, typecheck, test, lint.