| name | move-page |
| description | Move a blog page directory to a new location and update every internal link to it across the codebase. Use whenever the user asks to move, relocate, re-parent, or restructure a page ("move X under Y", "put X into Y", "move X out of A into B"). Handles both relative (`/path`) and absolute (`https://blog.mbedded.ninja/path`) link forms, and updates links to sub-pages too. |
Move Page Skill
Use this skill when the user asks to move a blog page from one location in the content tree to another. Pages live at src/content/pages/<...>/index.mdx; moving a page means moving the directory that contains its index.mdx.
Step 1: Confirm source and target
Determine:
- Source directory — the page's current directory, e.g.
src/content/pages/programming/general/checksums.
- Target directory — the new location, e.g.
src/content/pages/programming/checksums.
For "move X under Y", the target is <Y-dir>/<X-slug> — preserve the slug unless the user asks to rename.
If the user is moving multiple pages in one request (e.g. a parent page and a child page), do them one at a time, parent first. Otherwise the child move target may not exist yet.
Step 2: Find every link to the page
Run the bundled search script with the page's URL path (the path that appears inside markdown links like [text](/url/path)):
python .claude/skills/move-page/find-page-links.py /programming/general/checksums
The script:
- Walks every
.mdx and .md file under src/content/.
- Matches both relative (
/page-path) and absolute (https://blog.mbedded.ninja/page-path) link forms.
- Matches sub-pages too — searching for
/foo/bar also reports links to /foo/bar/baz.
- Stops short of unrelated slug suffixes —
/foo/bar does NOT match /foo/bar-other.
Output is file:line: matched-line for every hit. Read the output before moving so you have the full list of files to update in step 4.
Step 3: Move the directory with git mv
git mv src/content/pages/programming/general/checksums src/content/pages/programming/checksums
Use git mv (not plain mv) so the rename is tracked in git history.
The target's parent directory must already exist. Usually it does because the parent is itself an existing page directory (it has its own index.mdx). If the parent doesn't exist, that's a sign you should check the target with the user before continuing.
Step 4: Update every found link
For each match returned by step 2, use the Edit tool to rewrite the link. Both forms must be updated where they appear:
- Relative:
/old/path → /new/path
- Absolute:
https://blog.mbedded.ninja/old/path → https://blog.mbedded.ninja/new/path
Use Edit (not sed/awk). Each replacement should be visible and reviewable.
Step 5: Verify
Re-run the search script with the OLD path. It should report no matches. Then run git status and confirm:
- The directory rename appears as
renamed: lines.
- Every linking file appears as
modified: with the link rewritten.
Notes
- The script is stdlib-only — no
uv or extra packages required.
- Sub-page links update automatically because the search matches by prefix. After moving
/foo/bar to /baz/bar, a link to /foo/bar/child becomes /baz/bar/child with the same prefix replacement.
- Inbound links from other sites cannot be updated — only files inside this repo are touched.
- If the search returns dozens of hits, pause and confirm the move with the user — broad page moves create a lot of churn and may indicate the target slug isn't quite right.
