| name | fetch-post-manifest |
| description | Read the canonical list of posts from the claudecodeformarketers.com repo, normalize frontmatter, flag missing fields. Source of truth every other skill joins against. Run first in the weekly pipeline. |
fetch-post-manifest
Reads the Astro content collection at src/content/blog/ in the site repo, parses frontmatter from every .md / .mdx file, and writes a normalized posts.json to the snapshot store at reports/{week}/posts.json.
When to invoke
Run this skill first in the weekly pipeline. Every other fetcher and analyzer joins against posts.json by slug or url.
How to invoke
python .claude/skills/fetch-post-manifest/scripts/fetch.py [--week 2026-W18]
Defaults to the current ISO week if --week is omitted.
Inputs
SITE_URL — base URL of the site (e.g. https://claudecodeformarketers.com)SITE_REPO_LOCAL — local checkout path of the site repo (used in dev)SITE_REPO + GITHUB_TOKEN — used in the Routine to clone freshCONTENT_PATH — relative path to the content collection (default src/content/blog)
Output
Writes reports/{week}/posts.json with the schema below. Posts missing target_keyword or description are flagged in manifest_issues but not excluded — analyzers decide what to do with them.
{
"generated_at": "2026-05-04T06:00:00Z",
"week": "2026-W18",
"post_count": 27,
"posts": [
{
"slug": "what-is-vibe-marketing",
"url": "https://claudecodeformarketers.com/blog/what-is-vibe-marketing",
"title": "...",
"description": "...",
"pub_date": "2026-04-21",
"updated_date": null,
"target_keyword": null,
"secondary_keywords": [],
"categories": ["Concepts"],
"tags": ["workflows", "prompting"],
"draft": false,
"frontmatter_complete": false,
"missing_fields": ["target_keyword"]
}
],
"manifest_issues": [
{ "slug": "what-is-vibe-marketing", "issue": "missing target_keyword", "severity": "warn" }
]
}
Notes
- Drafts are included in the output but flagged via the
draft field. Downstream skills filter them out as needed.
- Slug is the filename without extension. URL assumes the Astro route
/blog/{slug} — adjust if the site changes.
target_keyword and secondary_keywords are not yet defined in the site's Zod schema. Until they are, every post will land in manifest_issues. That's expected; the report uses this to drive the frontmatter audit.
