# Changelog

> **This is the PUBLIC form of this artifact. Findings are listed in full, but the details of SECURITY findings — which rule fired, in which file, on which line, and how to fix it — are deliberately withheld, and any secret-scanner results are excluded entirely. Where detail is absent here it was REMOVED FOR PUBLICATION; it is not missing from the analysis. The complete artifact is available from the repository owner.**

## Score

- CAI 64 → 62 (-2.4)
- Rubric changed (rubric-2026.10.5 → rubric-2026.11.4) — scores are not directly comparable.

## Lenses

- Code Health 63 → 62 (-1.3)
- Architecture 86 → 87 (+1.8)
- Maturity 75 → 75 (+0.1)
- Readiness 57 → 57 (+0.0)
- Security 71 → 71 (+0.0)
- Accessibility 61 (new)
- Performance 100 → 100 (+0.0)

## Resolved (13)

- BaseSessionMixin.__generate_options__ (cognitive 32) (scrapling/engines/_browsers/_base.py)
- BaseSessionMixin.__generate_options__ (cyclomatic 20) (scrapling/engines/_browsers/_base.py)
- Documentation: contradicts the code (docs/requirements.txt)
- Documentation: no project overview (README.md)
- Duplicate configuration types. `PlaywrightConfig` and `_fetch_params` contain nearly identical fields (e.g., `timeout`, `wait`, `proxy`, `headers`), as do `StealthConfig` and `StealthFetchParams`. This duplication increases maintenance burden and risk of drift.
- Duplicated block (39–40 lines × 2) (scrapling/engines/_browsers/_stealth.py)
- Hotspot: scrapling/engines/_browsers/_base.py (scrapling/engines/_browsers/_base.py)
- Hotspot: scrapling/parser.py (scrapling/parser.py)
- Inconsistent naming for state persistence. The session objects use `save_state`/`load_state`, while the MCP server uses `browser_save_state`/`browser_load_state`. This inconsistency extends to other methods like `close` vs `close_session`.
- Near-duplicate member pair (69 shared lines) (scrapling/engines/_browsers/_stealth.py)
- No ADRs found
- Overlapping functionality between browser fetch methods. `browser_fetch` and `browser_fetch_once` appear to perform similar HTTP-like extraction tasks within a browser context, while `session_make_request` handles HTTP requests within a session. The distinction between 'fetch' and 'make_request' in the context of a browser session is unclear and likely redundant.
- Redundant HTTP verb methods vs generic fetch. The CLI exposes specific methods for GET, POST, PUT, and DELETE, but also exposes a generic `fetch` method. This creates confusion about which method to use for standard HTTP requests versus complex browser-based fetching.

## New (24)

- Duplicated block (41–42 lines × 2) (scrapling/engines/_browsers/_stealth.py)
- Edited copy of a member (115 corresponding lines) (scrapling/engines/_browsers/_stealth.py)
- Edited copy of a member (14 corresponding lines) (scrapling/engines/_browsers/_stealth.py)
- Edited copy of a member (14 corresponding lines) (scrapling/spiders/spider.py)
- Edited copy of a member (16 corresponding lines) (scrapling/engines/static.py)
- Edited copy of a member (22 corresponding lines) (scrapling/core/ai/server.py)
- Edited copy of a member (22 corresponding lines) (scrapling/engines/_browsers/_base.py)
- Edited copy of a member (24 corresponding lines) (scrapling/engines/toolbelt/navigation.py)
- Edited copy of a member (30 corresponding lines) (scrapling/engines/_browsers/_base.py)
- Edited copy of a member (38 corresponding lines) (scrapling/engines/toolbelt/convertor.py)
- Edited copy of a member (49 corresponding lines) (scrapling/fetchers/stealth_chrome.py)
- Edited copy of a member (7 corresponding lines) (scrapling/parser.py)
- Edited copy of a member (8 corresponding lines) (scrapling/engines/_browsers/_remote.py)
- Edited copy of a member (8 corresponding lines) (scrapling/spiders/templates/crawler.py)
- Edited copy of a member (89 corresponding lines) (scrapling/engines/toolbelt/convertor.py)
- Edited copy of a member (9 corresponding lines) (scrapling/engines/_browsers/_base.py)
- High: security finding (details withheld)
- Hotspot: scrapling/core/ai/_browser_actions.py (scrapling/core/ai/_browser_actions.py)
- Near-duplicate member pair (71 shared lines) (scrapling/engines/_browsers/_stealth.py)
- No ADRs found (pyproject.toml)
- …and 4 more

## Changes since last survey

- 44 commits — 31 feature/other, 13 fixes

## By area

- docs/locales — 8 commits
- (root) — 7 commits
- scrapling/core — 5 commits
- (repo) — 4 commits
- scrapling/parser.py — 3 commits
- scrapling/spiders — 3 commits
- tests/ai — 3 commits
- agent-skill/Scrapling-Skill — 2 commits
- scrapling/engines — 2 commits
- tests/fetchers — 2 commits
- tests/spiders — 2 commits
- .github/workflows — 1 commit
- images/olostep.png — 1 commit
- scrapling/cli.py — 1 commit

## Notable commits

- fix: fix(cli): make --ai-targeted enable ad blocking on fetch (#527)
- fix: fix(fetchers): decode percent-encoded credentials in proxy URLs (#545)
- fix: fix(parser): handle empty HTML roots (#536)
- fix: fix(parser): honor zero similarity threshold without reference attributes (#539)
- fix: fix(parser): return the value of XPath expressions that do not select nodes (#552)
- fix: fix(shell): accept curl's -L flag (#554)
- fix: fix(shell): stop dropping cookies whose values SimpleCookie rejects (#541)
- fix: fix(spiders): fetch robots.txt once per domain under concurrent requests (#538)
- fix: fix(spiders): handle CR record separators in CSV feeds (#551)
- fix: fix(spiders): isolate temporary files for concurrent cache writes (#546)
- fix: fix(spiders): skip malformed links during extraction (#548)
- fix: test: fix flaky Cloudflare readiness test (#515)
- fix: test: fix repeated issue on GitHub Actions
- change: Merge branch 'main' into dev
- change: Merge branch 'main' into dev
- change: Merge branch 'main' into dev
- change: agent: update skill with the current changes
- change: ci: add a new workflow to handle stale issues/PRs
- change: docs: Update changelog with the latest changes
- change: docs: add a new sponsor
- …and 24 more

## Architecture

- Unchanged — 0 containers · 1 contexts · 0 edges
