Add developer-index and research-index session skills

[Max17190]

Why

The developer and research index skills were sitting in firecrawl/skills, which is the build repo. They are session skills (answer a question now), so they belong next to firecrawl developer and firecrawl research.

Summary

• Add firecrawl-developer-index and firecrawl-research-index as standalone CLI skills, registered in the Claude plugin lists.
• Route paper and library/API questions from firecrawl / firecrawl-search to those skills instead of embedding the recipes in web search.
• Match live production shapes: developer hits use the id prefix for kind (no type field); search-papers hits carry paperId, primaryId, ids, title, abstract, and score.
• Companion change: those skills are removed from firecrawl/skills on Move index skills out of the build repo and align developer search with the live wire skills#11

Test Plan

• npx skills add firecrawl/cli@firecrawl-developer-index and @firecrawl-research-index install
• firecrawl developer "retry backoff" --limit 2 --json results have id, url, passages and no type
• firecrawl research search-papers "CRISPR base editing" --limit 2 --json results have paperId, primaryId, ids, title, abstract, score
• firecrawl developer --help no longer claims a type field
• firecrawl-search description no longer claims paper search as its job

---

Summary by cubic

Adds firecrawl-developer-index and firecrawl-research-index as standalone session skills and makes firecrawl-search web-only. This clarifies routing for coding and paper queries and aligns response shapes with production.

• Registers the two skills in .claude-plugin/marketplace.json and .claude-plugin/plugin.json, adds SKILL docs, and updates firecrawl-search/firecrawl docs to point to the new skills and use shorter triggers.
• Developer search behavior change: results drop the type field; infer kind from the id prefix (doc:, issue:, pull_request:, readme:). Required: update any parsers that read type; CLI help in src/index.ts now documents the prefix.
• Paper search shape is explicit: search-papers returns paperId, primaryId, ids, title, abstract, and score. Required: ensure clients read these fields.
• firecrawl-search no longer claims to hit the paper index; --categories research is a website filter only, and --categories developer just weighs developer hits in the same web call. Required: call firecrawl research/firecrawl-research-index for papers and firecrawl-developer-index for library/API/error/bug questions.
• Optional coordination: remove these skills from the build repo when this lands.

^{Written for commit 3c9202d. Summary will update on new commits.}

[cubic-dev-ai]

All reported issues were addressed across 7 files

<sub>Reply with feedback, questions, or to request a fix.

Fix all with cubic | Re-trigger cubic</sub>

[cubic-dev-ai]

All reported issues were addressed across 4 files (changes from recent commits).

<sub>Reply with feedback, questions, or to request a fix.

Fix all with cubic | Re-trigger cubic</sub>

[hmishra2250]

LGTM!