You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: AGENTS.md
+44-24Lines changed: 44 additions & 24 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -8,6 +8,19 @@ The ONLY reason you should stop is if you have a blocking question for the user.
8
8
9
9
Smart case-aware search and replace across code and files with atomic apply and undo.
10
10
11
+
## CRITICAL: Context-Specific CLAUDE.md Files
12
+
13
+
This repository contains multiple CLAUDE.md files throughout the codebase, each providing context-specific instructions for different parts of the project:
-**`/renamify-mcp/CLAUDE.md`** - MCP server specific rules (Ultracite coding standards)
17
+
-**`/renamify-vscode/CLAUDE.md`** - VS Code extension specific rules and patterns
18
+
-**Other subdirectories may have their own CLAUDE.md files**
19
+
20
+
**You MUST read the context-specific CLAUDE.md file whenever working in that part of the codebase.** These files contain critical rules and patterns specific to that component. For example, the MCP server has strict TypeScript/accessibility rules enforced by Ultracite that don't apply to Rust code.
21
+
22
+
**Note:**`/renamify-e2e-test/` is a temporary directory created during e2e tests (a clone of the entire repo for testing self-renaming). It's not part of the source tree and should be ignored.
23
+
11
24
## Project background
12
25
13
26
- Greenfield, unreleased. No backwards compatibility. Delete and tidy old code immediately.
@@ -210,12 +223,17 @@ The unrestricted levels (`-u` flag) control ignore behavior:
210
223
211
224
1. Begin next task and keep going
212
225
- Implement tasks in order unless dependencies dictate otherwise
213
-
2. After each task
214
-
- Run tests and linters
226
+
2. After each task - BEFORE marking as complete:
227
+
- Run `task ci` - this runs all linters, formatters, and tests that run on CI
228
+
- Run `coderabbit --plain` - get third-party code review
229
+
- Implement ALL suggestions from CodeRabbit
230
+
- Repeat `task ci` and `coderabbit --plain` until BOTH pass with no issues or suggestions
215
231
- Update docs if anything changed
216
232
- Commit with a clear message
217
233
3. Only pause for a blocking user question
218
234
235
+
**CRITICAL**: Never claim a task is complete without running `task ci` and `coderabbit --plain` successfully. This is non-negotiable.
236
+
219
237
## User preferences
220
238
221
239
- Keep responses short and terse in chats
@@ -312,36 +330,38 @@ This file is not set in stone, it is a living document that you should update as
312
330
313
331
## CRITICAL: ALWAYS TEST BEFORE CLAIMING COMPLETION
314
332
315
-
**NEVER claim something is done without running tests and linters for the language you're working in:**
333
+
**NEVER claim something is done without completing the full workflow:**
316
334
317
-
For JavaScript/TypeScript projects:
335
+
1.**Run `task ci`** - This single command runs ALL linters, formatters, and tests across all languages
336
+
- Rust: cargo fmt, clippy, test
337
+
- TypeScript/JavaScript: pnpm check, pnpm test (for all TS projects)
338
+
- Docs: prettier, biome
339
+
- Shell scripts: shellcheck
340
+
- TOML: taplo
318
341
319
-
-`pnpm check` or `npm run lint` - Run linting and type checking
320
-
-`pnpm test` or `npm test` - Run all tests
321
-
-`pnpm build` or `npm run build` - Ensure the project builds
342
+
2.**Run `coderabbit --plain`** - Get third-party AI code review
343
+
- Implement ALL suggestions from CodeRabbit
344
+
- Re-run `task ci` after implementing suggestions
345
+
- Repeat until CodeRabbit returns no further suggestions
322
346
323
-
For Rust projects:
347
+
3.**Only then** can you claim the task is complete
324
348
325
-
-`cargo build` - Ensure the project compiles
326
-
-`cargo clippy` - Run linter (warnings should be treated as errors)
327
-
-`cargo test` - Run all tests
328
-
-`cargo fmt --check` - Check formatting
349
+
**NO EXCEPTIONS.** This is the workflow. Follow it religiously.
0 commit comments