fix(scan): drop requirements.txt from publish, bump to 2.1.3

The Tank registry scanner reads requirements.txt and was flagging
PYSEC-2024-277 on joblib (already patched in 1.4.2+; we ship 1.5.3)
and a typosquat false-positive ('hpack' vs 'black'). hpack is the
legit HPACK HTTP/2 header library, a transitive dep of h2.

Both findings are spurious, but the easiest fix is to not ship
requirements.txt inside the package tarball — the deps live in
pyproject.toml, which is what uvx and pip use at install time.
The MCP runtime does not need requirements.txt.

Also updated the README to point users at 'pip install -e .' instead
of 'pip install -r requirements.txt' for consistency.

Author: elad12390

README.md

@@ -111,7 +111,7 @@ uv sync              # creates the virtual environment
 uv run crawl4ai-setup  # installs Chromium for crawl4ai
 ```
 
-> You can also use `pip install -r requirements.txt` if you prefer pip over uv.
+> You can also use `pip install -e .` if you prefer pip over uv (deps come from `pyproject.toml`).
 
 ## Installation
 

tank.json

@@ -1,6 +1,6 @@
 {
   "name": "@elad12390/web-research-assistant",
-  "version": "2.1.2",
+  "version": "2.1.3",
   "description": "MCP server for web research with 14 tools: SearXNG + Exa AI search, URL crawling, stealth scraping (Cloudflare bypass), package registry lookup (npm/PyPI/crates), GitHub stats, changelog fetching, tech comparison, error translation, API docs discovery, stock image search, and service health. Triggers: web search, find packages, crawl url, scrape website, translate error, compare frameworks, stock images, service status, search github, api docs.",
   "permissions": {
     "network": {
@@ -17,7 +17,6 @@
     "files": [
       "src/**",
       "pyproject.toml",
-      "requirements.txt",
       "SKILL.md",
       "tank.json",
       "README.md"