Add ERT tests, CI workflow, and shared command helpers

[chocoelho]

Summary by Sourcery

Add automated testing and CI for the cursor-agent Emacs package and refactor command execution helpers for better ANSI-aware output handling.

New Features:

• Introduce ERT-based unit tests for cursor-agent in cursor-agent-test.el, including configuration, helper, and command presence checks.

Enhancements:

• Refactor repeated shell/compilation invocations into shared helpers that run commands in compilation buffers with ANSI color handling and optional installation checks.
• Improve robustness of setup and status checks by guarding calls to authentication functions and avoiding unused lambda parameters.
• Document how to run the test suite locally and note that tests run automatically in CI.

CI:

• Add a GitHub Actions workflow to lint, byte-compile, and run ERT tests for the package on pushes, pull requests, and manual triggers.

[sourcery-ai]

Reviewer's Guide

Adds an ERT-based test suite and GitHub Actions CI for the cursor-agent Emacs package, introduces shared helpers for running CLI commands in compilation buffers with ANSI color handling and installation checks, and refines readiness/auth checks and docs to support the new workflow.

Class diagram for cursor-agent command helpers and list commands

classDiagram
    class CursorAgentModule {
      +cursor-agent-interactive(initial-prompt)
      +cursor-agent-process-region(start end)
      +cursor-agent-resume(chat-id)
      +cursor-agent-list-sessions()
      +cursor-agent-status()
      +cursor-agent-list-models()
      +cursor-agent-mcp-list()
      +cursor-agent-login()
      +cursor-agent-installed-p()
      +cursor-agent-ready-p()
      +cursor-agent-open-readme()
      -cursor-agent--run-command-in-compilation-buffer(command buffer-name header-text)
      -cursor-agent--run-list-command(command buffer-name header-text)
      -cursor-agent--install-sentinel(buffer-name)
    }

    class CursorAgentCLI {
      +ls
      +status
      +models
      +mcp_list
      +login
    }

    CursorAgentModule ..> CursorAgentCLI : uses

    CursorAgentModule ..> CursorAgentModule : cursor-agent--run-list-command uses cursor-agent-installed-p
    CursorAgentModule ..> CursorAgentModule : cursor-agent--run-list-command uses cursor-agent--run-command-in-compilation-buffer

    CursorAgentModule ..> CursorAgentModule : cursor-agent-list-sessions uses cursor-agent--run-list-command
    CursorAgentModule ..> CursorAgentModule : cursor-agent-status uses cursor-agent--run-list-command
    CursorAgentModule ..> CursorAgentModule : cursor-agent-list-models uses cursor-agent--run-list-command
    CursorAgentModule ..> CursorAgentModule : cursor-agent-mcp-list uses cursor-agent--run-list-command

    CursorAgentModule ..> CursorAgentModule : cursor-agent-interactive uses cursor-agent--run-command-in-compilation-buffer
    CursorAgentModule ..> CursorAgentModule : cursor-agent-process-region uses cursor-agent--run-command-in-compilation-buffer
    CursorAgentModule ..> CursorAgentModule : cursor-agent-ready-p uses cursor-agent-installed-p

Loading

File-Level Changes

Change	Details	Files
Introduce shared helpers for running cursor-agent shell commands in compilation buffers with ANSI color handling and reuse them across interactive commands.	Add cursor-agent--run-command-in-compilation-buffer to centralize async-shell-command usage with compilation-mode and compilation-filter for ANSI colors. Add cursor-agent--run-list-command to wrap list-style CLI invocations with an installation check and shared buffer setup. Refactor cursor-agent-prompt and cursor-agent-region to use the new helper instead of duplicating compilation buffer setup logic. Refactor cursor-agent-list-sessions, cursor-agent-status, cursor-agent-list-models, and cursor-agent-mcp-list to use cursor-agent--run-list-command and update their docstrings to document ANSI color handling. Adjust cursor-agent-login to set the process filter to compilation-filter so its compilation buffer also handles ANSI colors.	cursor-agent.el
Harden installation/authentication flow and minor internals for better robustness and tooling cleanliness.	Update cursor-agent-check-readiness to guard cursor-agent-check-auth with fboundp so readiness checks don’t error when the function is unavailable. Ignore the process argument in cursor-agent--install-sentinel using a _proc parameter to avoid unused variable warnings. Tidy control flow/paren structure in cursor-agent-check-auth and cursor-agent-readme without changing behavior.	cursor-agent.el
Add an ERT-based test suite that validates installation detection, authentication checks, configuration defaults, helper utilities, and command availability, plus an interactive test runner.	Create cursor-agent-test.el with unit tests for cursor-agent-installed-p and cursor-agent-check-auth using cl-letf-based mocks. Add tests for configuration defaults (cursor-agent-command, cursor-agent-default-output-format, cursor-agent-use-force). Add tests for helper functions such as cursor-agent--local-bin-in-path-p and cursor-agent--get-path-instruction for bash/zsh. Add loading/bootstrapping helpers to robustly require or load cursor-agent.el in varying environments, including force-loading specific defuns if needed. Add integration-style tests to assert that all expected interactive commands are defined and commandp, plus an interactive cursor-agent-test-run-all test runner. Provide the cursor-agent-test feature for easy loading from Emacs and CI.	cursor-agent-test.el
Document how to run the new test suite locally and from the command line.	Add a Testing section to README.md describing how to load cursor-agent-test.el and run ERT tests interactively via M-x ert. Document a non-interactive CLI invocation using emacs --batch with ert-run-tests-batch-and-exit, suitable for local scripting or CI. Mention that tests run automatically in GitHub Actions.	README.md
Add a GitHub Actions workflow to lint, byte-compile, and run the ERT test suite on pushes, pull requests, and manual triggers.	Create .github/workflows/test.yml with a Test job running on push, pull_request, and workflow_dispatch. Use purcell/setup-emacs to install Emacs 29.1 on ubuntu-latest. Run leotaku/elisp-check with the melpa check on cursor-agent.el for linting and byte compilation. Run leotaku/elisp-check with the ert check on cursor-agent-test.el to execute the ERT suite in CI.	.github/workflows/test.yml

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[sourcery-ai]

Hey - I've found 2 issues, and left some high level feedback:

• The four commands that list sessions/status/models/MCP now share nearly identical logic for setting up the buffer, mode, and async shell command; consider factoring this into a small helper to avoid duplication and keep future changes to the behavior in one place.
• The new safety check in cursor-agent-verify-setup uses fboundp on cursor-agent-installed-p; if this function is guaranteed to exist whenever cursor-agent is loaded, you might prefer failing earlier via a require or simplifying the check to avoid raising a runtime error in normal usage.
• In the tests, mocking core functions like executable-find and shell-command-to-string with fset can have surprising global side effects (especially under parallel test execution); consider using cl-letf or letf to temporarily rebind these functions in a more scoped way.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- The four commands that list sessions/status/models/MCP now share nearly identical logic for setting up the buffer, mode, and async shell command; consider factoring this into a small helper to avoid duplication and keep future changes to the behavior in one place.
- The new safety check in `cursor-agent-verify-setup` uses `fboundp` on `cursor-agent-installed-p`; if this function is guaranteed to exist whenever `cursor-agent` is loaded, you might prefer failing earlier via a `require` or simplifying the check to avoid raising a runtime error in normal usage.
- In the tests, mocking core functions like `executable-find` and `shell-command-to-string` with `fset` can have surprising global side effects (especially under parallel test execution); consider using `cl-letf` or `letf` to temporarily rebind these functions in a more scoped way.

## Individual Comments

### Comment 1
<location> `cursor-agent.el:459-461` </location>
<code_context>
+          (with-current-buffer (get-buffer-create buffer-name)
+            (erase-buffer)
+            (insert "Test content")
+            (compilation-mode))
+          (should (buffer-live-p (get-buffer buffer-name)))
+          (should (eq major-mode 'compilation-mode)))
</code_context>

<issue_to_address>
**issue (bug_risk):** Using `compilation-mode` with `async-shell-command` likely won’t run `compilation-filter-hook`, so ANSI color handling may not work as expected.

Because `async-shell-command` installs `shell-command--command-output-filter` instead of `compilation-filter`, `compilation-filter-hook` is never run and ANSI escape sequences will not be processed. To ensure `compilation-mode`’s hooks (including ANSI handling) actually apply, consider using `compilation-start` instead of `async-shell-command`, or explicitly set the process filter to `compilation-filter`. The same pattern shows up in the status/models/MCP commands and likely needs a shared fix.
</issue_to_address>

### Comment 2
<location> `README.md:353` </location>
<code_context>
+M-x ert RET cursor-agent-test-.* RET
+```
+
+Or from command line:
+
+```bash
</code_context>

<issue_to_address>
**nitpick (typo):** Consider changing "from command line" to "from the command line" for grammatical correctness.

Rephrase this line to: `Or from the command line:` to improve grammar and readability.

```suggestion
Or from the command line:
```
</issue_to_address>

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

[chocoelho]

@sourcery-ai summary

[chocoelho]

@sourcery-ai review

[chocoelho]

@sourcery-ai title

[sourcery-ai]

Hey - I've found 3 issues, and left some high level feedback:

• For cursor-agent-login, consider reusing cursor-agent--run-command-in-compilation-buffer to avoid duplicating the async command + compilation-filter setup and keep behavior consistent with the other command entry points.
• The switch of the list-style commands (cursor-agent-list-sessions, cursor-agent-status, cursor-agent-list-models, cursor-agent-mcp-list) from synchronous shell-command-to-string with special-mode to asynchronous compilation-mode changes both timing and buffer keybindings; it may be worth confirming this UX change is intentional or gating it behind a user option.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- For `cursor-agent-login`, consider reusing `cursor-agent--run-command-in-compilation-buffer` to avoid duplicating the async command + `compilation-filter` setup and keep behavior consistent with the other command entry points.
- The switch of the list-style commands (`cursor-agent-list-sessions`, `cursor-agent-status`, `cursor-agent-list-models`, `cursor-agent-mcp-list`) from synchronous `shell-command-to-string` with `special-mode` to asynchronous `compilation-mode` changes both timing and buffer keybindings; it may be worth confirming this UX change is intentional or gating it behind a user option.

## Individual Comments

### Comment 1
<location> `cursor-agent.el:132-141` </location>
<code_context>
+      (setq cursor-agent-authenticated-p nil)
+      nil))))
+
+(defun cursor-agent--run-command-in-compilation-buffer (command buffer-name header-text)
+  "Run COMMAND in a compilation buffer with proper ANSI color handling.
+COMMAND is the shell command to execute.
+BUFFER-NAME is the name of the buffer to use.
+HEADER-TEXT is the text to insert at the start of the buffer.
+Sets the process filter to compilation-filter to ensure ANSI colors are handled."
+  (let ((buffer (get-buffer-create buffer-name)))
+    (with-current-buffer buffer
+      (erase-buffer)
+      (when header-text
+        (insert header-text))
+      (compilation-mode))
+    ;; Use async-shell-command but set the filter to compilation-filter
+    ;; to ensure compilation-filter-hook runs, which handles ANSI color codes
+    (let ((process (async-shell-command command buffer-name)))
+      (when process
+        (set-process-filter process 'compilation-filter)))
</code_context>

<issue_to_address>
**suggestion:** Consider reusing this helper in `cursor-agent-login` to avoid duplicating process/filter setup logic.

The same process start + `compilation-mode` + `set-process-filter` sequence exists in `cursor-agent-login`. Calling `cursor-agent--run-command-in-compilation-buffer` from there (after any login-specific preamble) would centralize the async/ANSI handling and keep future behavior changes in one place.

Suggested implementation:

```
(defun cursor-agent-login ()
  "Login to Cursor Agent."
  (interactive)
  (let* ((buffer-name "*Cursor Agent Login*")
         (header-text (concat
                       "Cursor Agent Login\n\n"
                       "Follow the instructions in this buffer to log in.\n\n")))
    (cursor-agent--run-command-in-compilation-buffer
     "cursor-agent login"
     buffer-name
     header-text)))

```

1. If the actual implementation of `cursor-agent-login` differs from the SEARCH block above (e.g., different buffer name, different header text, extra checks like `cursor-agent-installed-p` or environment validation), adjust the replacement to preserve that existing behavior while delegating the process setup to `cursor-agent--run-command-in-compilation-buffer`.
2. Ensure that any login-specific preamble (such as additional `insert` calls, instructions, or state setup) is incorporated into the `header-text` argument before calling `cursor-agent--run-command-in-compilation-buffer`, or executed just before that call in the function body.
3. If `cursor-agent-login` previously used a different shell command (e.g., included flags or environment variables), update the `"cursor-agent login"` string accordingly to match the original command.
</issue_to_address>

### Comment 2
<location> `cursor-agent.el:151-160` </location>
<code_context>
+        (set-process-filter process 'compilation-filter)))
+    (pop-to-buffer buffer-name)))
+
+(defun cursor-agent--run-list-command (command buffer-name header-text)
+  "Helper function to run a list command with installation check.
+COMMAND is the shell command to execute.
+BUFFER-NAME is the name of the buffer to use.
+HEADER-TEXT is the text to insert at the start of the buffer.
+Checks if cursor-agent is installed before running the command."
+  (unless (cursor-agent-installed-p)
+    (if (y-or-n-p "Cursor Agent CLI not found. Would you like to install it now? ")
+        (cursor-agent-install)
+      (user-error "Cursor Agent CLI not found.  Run 'M-x cursor-agent-install' to install")))
+  (cursor-agent--run-command-in-compilation-buffer
+   command
+   buffer-name
+   header-text))

 (defun cursor-agent--local-bin-in-path-p ()
</code_context>

<issue_to_address>
**question:** Listing commands no longer put buffers into `special-mode`, which changes interaction semantics.

These commands now yield buffers in `compilation-mode` instead of `special-mode`, so they’ve lost the previous read-only/navigation/quit behavior. If ANSI handling is the main reason for this change, consider re-enabling a read-only mode (e.g. `special-mode`) after the process completes, or explicitly marking the buffer read-only, so the prior interaction model is preserved while still using `compilation-filter`.
</issue_to_address>

### Comment 3
<location> `cursor-agent.el:609` </location>
<code_context>
           (view-mode)
           (message "Press 'q' to quit view mode"))
-      (user-error "README.md not found.  Please ensure the package is properly installed"))))
+      (user-error "README.md not found.  Please ensure the package is properly installed")))))

 (provide 'cursor-agent)
</code_context>

<issue_to_address>
**issue (bug_risk):** There appears to be one extra closing parenthesis at the end of `cursor-agent-open-readme`.

This extra `)` will unbalance the surrounding forms and prevent the file from reading/byte-compiling. The closing parens here should match the previous version (remove one `)`).
</issue_to_address>

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

[sourcery-ai]

issue (bug_risk): There appears to be one extra closing parenthesis at the end of cursor-agent-open-readme.

This extra ) will unbalance the surrounding forms and prevent the file from reading/byte-compiling. The closing parens here should match the previous version (remove one )).