feat(profile-edit): add 13 tools to edit every profile section

[Gabrcodes]

Summary

13 new tools for editing the authenticated user's LinkedIn profile. All tools use LinkedIn's overlay edit forms, fill fields by their label text (no fragile CSS class selectors), and are annotated destructiveHint.

Tool	What it edits
edit_profile_intro	First name, last name, headline, location, industry
edit_profile_about	About/Summary section text
add_experience	Work experience entries
add_education	Education entries
add_skill	Skills (with typeahead suggestion support)
add_certification	Certifications with credential ID/URL
add_volunteer_experience	Volunteer work
add_project	Projects with URL
add_publication	Publications
add_course	Courses
add_language	Languages with proficiency level
add_honor	Honors and awards

Implementation notes

• _fill_field_by_label uses Object.getOwnPropertyDescriptor with the correct prototype per element type (HTMLTextAreaElement, HTMLSelectElement, or HTMLInputElement) so React observes changes
• _select_dropdown_by_label uses the same native setter pattern for <select> elements
• _resolve_my_username() navigates to /in/me/ once and caches the result for all edit methods in the same call

Changes

• scraping/extractor.py — all edit helper methods and _resolve_my_username
• tools/profile_edit.py — new file with 13 tool registrations
• server.py — register profile edit tools

Test plan

• 357 passed, 5 skipped, 0 failures
• ruff check and ruff format pass
• Live test: edit headline via edit_profile_intro
• Live test: add a skill via add_skill

[greptile-apps]

Greptile Summary

This PR adds 12 new MCP tools for editing the authenticated user's LinkedIn profile. All prior review concerns have been resolved: edit_profile_about now targets contenteditable elements, industry correctly waits for and clicks a typeahead suggestion, _resolve_my_username is instance-cached, the required-field guard in _edit_profile_section_entry now aborts on any missing required field, and the get_my_profile dead code has been removed. The implementation consistently avoids LinkedIn-specific CSS class names and relies on label text and standard HTML attributes throughout.

Confidence Score: 5/5

Safe to merge; all prior P1 concerns are addressed and the only remaining finding is a minor docstring count mismatch.

All previously flagged P1 issues (contenteditable about editor, industry typeahead, username caching, partial-save guard, dead code) are resolved in this revision. The sole new finding is a P2 documentation discrepancy (12 tools labelled as 13) that has no runtime impact.

No files require special attention beyond the minor docstring fix in tests/test_tools.py.

Important Files Changed

Filename	Overview
linkedin_mcp_server/scraping/extractor.py	Adds 12 edit helper methods and 4 private helpers (_fill_field_by_label, _select_dropdown_by_label, _click_save_in_dialog, _resolve_my_username with caching); all previous review concerns (contenteditable about, industry typeahead, username caching, required-field guard, dead code) are addressed
linkedin_mcp_server/tools/profile_edit.py	New file registering 12 MCP tools for profile editing; follows existing patterns for auth handling, progress reporting, and destructiveHint annotation
linkedin_mcp_server/server.py	Adds import and registration call for profile_edit tools; minimal, correct change
tests/test_tools.py	Adds TestProfileEditTool class with 12 happy-path tests and a shared helper; docstring incorrectly says '13 profile editing tools' when 12 are implemented
linkedin_mcp_server/tools/init.py	Updates module docstring to list the new profile-edit and connections tool categories

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    T["MCP Tool call\n(profile_edit.py)"] --> GE["get_ready_extractor()"]
    GE --> RU["_resolve_my_username()\n(cached in _my_username_cache)"]
    RU --> NAV["_navigate_to_page(URL)\nLinkedIn overlay form"]
    NAV --> WAIT["wait_for_selector\n(dialog / main form)"]
    WAIT -->|timeout| FAIL["return edit_failed"]
    WAIT --> FILL["_fill_field_by_label()\nJS: label → input/textarea/select\nnative React setter + events"]
    FILL --> DROP["_select_dropdown_by_label()\nJS: label → select\noption text search"]
    DROP --> REQ{required_fields\nall filled?}
    REQ -->|no| ABORT["return edit_failed\nmissing fields listed"]
    REQ -->|yes| SAVE["_click_save_in_dialog()\nSave / Apply / Done button"]
    SAVE -->|found| OK["return saved"]
    SAVE -->|not found| SF["return save_failed"]

Loading

Prompt To Fix All With AI

This is a comment left during a code review.
Path: tests/test_tools.py
Line: 749

Comment:
**Tool count mismatch: 12 tools, not 13**

The PR title, PR description, and this docstring all say "13" tools, but exactly 12 are implemented and registered (confirmed by the 12 mock methods in `_make_mock_extractor` and the 12 tool functions in `profile_edit.py`).

```suggestion
    """Tests for the 12 profile editing tools."""
```

How can I resolve this? If you propose a fix, please make it concise.

_{Reviews (11): Last reviewed commit: "fix: remove redundant case-insensitive s..." | Re-trigger Greptile}