feat: add comprehensive blog knowledge base and structured post index for enhanced querying and navigation

Author: JuanLara18

CLAUDE.md

@@ -0,0 +1,48 @@
+# Portfolio Repository
+
+Personal portfolio + technical blog (~90 posts on RAG, agents, knowledge graphs, data engineering, ML internals, math curiosities). React app under `front/`, blog content under `front/public/blog/posts/`.
+
+## Knowledge base for blog content
+
+When the user asks about blog topics — what they have written about, what to read on a topic, where to cite their own work, what concepts a post covers, prerequisites between posts — use the knowledge base in `knowledge-base/`.
+
+**Always start with `knowledge-base/KNOWLEDGE_BASE.md`.** It contains the manifest, reading paths, cross-cutting views (by stack / audience / depth / format), author context, and the full post catalog (slug + 1-line + concepts for every post).
+
+**For structured queries, use `knowledge-base/posts.json` via `jq`.** Examples:
+
+```bash
+# What does the blog cover about a concept?
+jq '.concept_index["rag"]' knowledge-base/posts.json
+
+# Posts that use a specific tech?
+jq '.tech_index["neo4j"]' knowledge-base/posts.json
+
+# Full metadata for one post (concepts, prereqs, teaches, depth, audio):
+jq '.posts["query-routing-agent-decisions"]' knowledge-base/posts.json
+
+# What posts depend on this one?
+jq '.prereq_graph.edges[] | select(.from=="model-context-protocol")' knowledge-base/posts.json
+
+# What should I read before this post?
+jq '.posts["ontology-to-agent-toolbox"].prereqs' knowledge-base/posts.json
+```
+
+`posts.json` is ~250K — never `cat` it whole, always `jq` to extract the section you need.
+
+**To read the full text of a post**, open the file at `front/public/blog/posts/<category>/<slug>.md`. Slugs are stable canonical IDs.
+
+When citing a post to the user, link as `https://juanlara18.github.io/portfolio/#/blog/<slug>`.
+
+## Build pipeline
+
+Blog content flows through `npm run prebuild` (in `front/`):
+
+1. `build-blog-data` — scans `posts/*.md`, parses frontmatter, writes `front/src/data/blogData.json`.
+2. `build-knowledge-base` — reads `blogData.json` + the YAML augmentation block in `KNOWLEDGE_BASE.md`, regenerates `knowledge-base/posts.json` and re-injects the auto-catalog into `KNOWLEDGE_BASE.md` (between `<!-- AUTO-CATALOG:* -->` markers).
+3. `build-sitemap`, `build-rss` — site metadata.
+
+To enrich a post's metadata (concepts, prereqs, teaches, tech, depth), edit the YAML block under `## Augmentation` in `KNOWLEDGE_BASE.md` and re-run `npm run build-knowledge-base`. Posts without an augmentation entry get sensible defaults (concepts ← tags, depth ← word count).
+
+## Authoring posts
+
+Use the `new-blog-post` skill for the full workflow (frontmatter, header image, audio generation). The post becomes available to the knowledge base on the next build.