Audit follow-ups: honest engines badge, dated live checks, canary env report

.github/env_report.py

@@ -0,0 +1,45 @@
+#!/usr/bin/env python3
+"""Print the versions a canary run actually had: `env_report.py [--chromium]`.
+
+Why it exists: when the canary goes red, the first question is whether the
+SITE moved or the runner did (Python, the TLS library, the driver, the
+browser). This answers it from the log of the run itself. It prints versions
+and the kind of endpoint, never an endpoint: a Scraping Browser URL carries
+its credentials in the login, and a log is a file people share.
+
+`--chromium` also launches the local Chromium to read its version, for the
+job that drives one; the Scraping Browser job has no local browser to ask.
+"""
+import os
+import platform
+import ssl
+import sys
+
+
+def version_of(dist):
+    try:
+        from importlib.metadata import version
+        return version(dist)
+    except Exception:
+        return "not installed"
+
+
+print("python      %s" % platform.python_version())
+print("platform    %s" % platform.platform())
+print("tls library %s" % ssl.OPENSSL_VERSION)
+for dist in ("requests", "urllib3", "beautifulsoup4", "playwright"):
+    print("%-11s %s" % (dist, version_of(dist)))
+print("endpoint    %s" % ("remote CDP (Scraping Browser)"
+                          if os.environ.get("MAKEMYTRIP_CDP_ENDPOINT")
+                          else "local browser, no endpoint"))
+
+if "--chromium" in sys.argv:
+    try:
+        from playwright.sync_api import sync_playwright
+        with sync_playwright() as p:
+            browser = p.chromium.launch()
+            print("chromium    %s" % browser.version)
+            browser.close()
+    except Exception as exc:
+        # A report must not fail the job it is describing.
+        print("chromium    could not be launched: %s" % type(exc).__name__)

.github/workflows/canary.yml

@@ -58,6 +58,8 @@ jobs:
           python -m pip install --upgrade pip
           pip install -r requirements.txt -r requirements-playwright.txt
           playwright install --with-deps chromium
+      - name: Environment (so a red run says whether the site or the runner moved)
+        run: python .github/env_report.py --chromium
       - name: Run
         id: run
         run: |
@@ -113,6 +115,8 @@ jobs:
         run: |
           python -m pip install --upgrade pip
           pip install -r requirements.txt -r requirements-playwright.txt
+      - name: Environment (so a red run says whether the site or the runner moved)
+        run: python .github/env_report.py
       # >= 3 pages, because with one page pagination is never exercised.
       - name: Run
         id: run

CHANGELOG.md

@@ -83,6 +83,19 @@ said so at the top of its release notes.
 - Comments, docs and the suite state their reasons in this repo's own words
   instead of pointing at section numbers in notes the public cannot read. A
   suite check keeps it that way.
+- The README's `engines` badge no longer lists the three engines as equals:
+  it says pyppeteer is measured over CDP and Selenium locally only, which is
+  what the engine table below it already said.
+- The README dates the last live checks beside the measurements table.
+
+### Added
+
+- `.github/env_report.py`, run by both canary jobs before the scrape: it
+  prints the Python, TLS library, driver and (for the local-browser job)
+  Chromium versions, and the kind of endpoint, never the endpoint. A red
+  canary then says from its own log whether the site or the runner moved.
+  Prompted by a third-party audit that met a TLS-library warning on an
+  older Python and could not tell it from the block.
 
 ## [0.1.1] — 2026-09-25
 

README.md

@@ -5,7 +5,7 @@
 [![canary](https://github.com/2scraper/makemytrip-scraper/actions/workflows/canary.yml/badge.svg)](https://github.com/2scraper/makemytrip-scraper/actions/workflows/canary.yml)
 ![python](https://img.shields.io/badge/python-3.9%20%7C%203.12-blue)
 [![licence](https://img.shields.io/badge/licence-MIT-green)](LICENSE)
-![engines](https://img.shields.io/badge/engines-playwright%20%7C%20selenium%20%7C%20pyppeteer-lightgrey)
+![engines](https://img.shields.io/badge/engines-playwright%20%7C%20pyppeteer%20via%20CDP%20%7C%20selenium%20local%20only-lightgrey)
 ![account](https://img.shields.io/badge/makemytrip%20account-not%20needed-brightgreen)
 
 Scrapes [makemytrip.com](https://www.makemytrip.com)'s **hotel listings**
@@ -117,6 +117,14 @@ Measured through a `country-in` Scraping Browser profile, 2026-09-24:
 | Goa `--url`, `--sort rating` | Playwright | 30 | 4.9, 5.0, 4.9, 4.8, 4.8 at the top |
 | Goa, one page | Scraper API + `cdpurl` | 5 | the page's server-rendered five; $0.0005 |
 
+**Last live checks on file** (this line goes stale; the dates are the point):
+a served three-page run through the Scraping Browser, 90 of 90 priced,
+2026-09-24 (the canary's own first run); pyppeteer through residential
+proxies, no rows, 2026-09-25; and from a datacentre address with no
+endpoint, the run reported Akamai's refusal as exit 3 and wrote nothing,
+2026-10-05, on Python 3.13 (the `datacentre` canary job's claim, re-checked
+by hand). No served run has been measured since 2026-09-24.
+
 ---
 
 ## Five things about MakeMyTrip that will look like bugs