CDP / navigation:
- Route all Page.navigate through process_with_interception so the
parent dispatcher stays responsive during long navs (fixes
Target.createTarget timeouts after a handful of newPage calls).
- Default waitUntil to domcontentloaded. Puppeteer and Playwright
don't pass it server-side and wait client-side via lifecycleEvent,
so the old Load default blocked nearly every nav on slow sites.
- 30s navigation hard ceiling, OBSCURA_NAV_TIMEOUT_MS to override.
- execute_scripts loop now checks wall-clock deadline per iteration.
- Emit Network.requestWillBeSent before Page.frameNavigated so
Playwright's FrameManager attaches the navigation response (#189).
- Use loaderId as the Document request id so page.goto() resolves
to a Response instead of null (#189).
- Fresh executionContextId per navigation, no more 'Cannot find
context with specified id' after the first nav (#192).
- POST navigation params plumbed through the intercept path.
DOM / Web API surface:
- MutationObserver: subtree filter walks parent chain, fires on
textContent / CharacterData.data / innerHTML, type filter respected,
oldValue passed.
- IntersectionObserver: multi-fire (initial + 120/500/1500ms),
disconnect / unobserve / takeRecords, root / rootMargin / thresholds.
- ResizeObserver: per-target firing from getBoundingClientRect, drop
the no-op override that was clobbering the real class.
- crypto.subtle.digest: real SHA-1/256/384/512 via op_subtle_digest.
- DOMParser / document.implementation.createHTMLDocument /
createDocument: detached doc with body, head, documentElement,
querySelector*, getElementById, getElementsByTagName.
- WebSocket / EventSource / BroadcastChannel: real EventTarget,
open fires on microtask.
- indexedDB: request, object store, transaction, database with
async onsuccess and in-memory Map storage.
- getComputedStyle: width/height/top/left from bbox so React
virtualization libs render items.
Site compatibility went from 3/40 to 175/175 reachable sites in
the sweep. Perf is equal or faster than pristine main on every
site tested (DCL default removes the block on long-tail subresources).
28/28 integration tests pass.
Picks up @Timtech4u's PR #33, which has been waiting since April. The
design (CLI flag + threading through BrowserContext + http client) is
theirs and is correct; this is a clean rewrite on top of current main
because repeated rebase rounds had been reverting unrelated work that
landed in the meantime.
Issue #4 blocked all loopback / RFC1918 / link-local fetches to prevent
SSRF. The only escape was OBSCURA_ALLOW_PRIVATE_NETWORK=1, which is
process-wide and clunky in pipelines. Multiple users (@Timtech4u in
the original PR, then @duartemvix and @necrogami in the comment
thread) reported they cannot use obscura for development against
http://localhost:N without it.
Adds --allow-private-network as a global CLI flag, works on fetch,
serve, and scrape. Threads through BrowserContext::with_storage_and_network
and the http client's validate_url. CdpContext gets a new_full
builder and obscura_cdp::start_with_full_serve_options entry point so
'obscura serve --allow-private-network' lets Puppeteer/Playwright
navigate to localhost too. Default remains blocked.
Backward compat: the env var still works (OR'd with the flag inside
validate_url). All existing constructors and entry points unchanged.
Verified end-to-end against a local http.server on 127.0.0.1: fetch
and CDP Page.navigate both blocked by default, both succeed with the
flag. Public URLs unaffected. Integration suite 28/28, full workspace
cargo test all green.
V8 embedded defaults to ~1.4 GB max-old-space on 64-bit. Pages with
heavy fingerprinting or analytics bundles (e.g. demo.fingerprint.com/
web-scraping) blow past it and the worker dies with
'Fatal JavaScript out of memory' and SIGTRAP, taking the CDP session
with it.
Always set --max-old-space-size=4096 (1024 on 32-bit targets) before
V8 startup, matching headless Chrome's default. User-supplied
--v8-flags are appended after the default so they win (V8 parses
left-to-right and the last value wins), which keeps the existing
escape hatch intact for memory-constrained environments.
Page.createIsolatedWorld and do_navigate both hardcoded context id 100
for every isolated world emission. After the second navigation Playwright
expected a fresh incrementing id (101, 102) while obscura kept re-using
100. Runtime.evaluate then failed with 'Cannot find context with
specified id: 101' because the id was never registered in
valid_context_ids and the client's bookkeeping had diverged from the
server's.
Added next_isolated_context_id counter on CdpContext (starts at 100) and
a next_isolated_context() helper that claims an id, registers it in
valid_context_ids, and increments. Both call sites now use it.
Two correlation invariants real Chrome holds were missing, so the
nav Response was never attached to goto().
requestId === loaderId for the main Document event. Both clients
gate isNavigationRequest() on this; we were emitting a synthetic
per-page id.
Network.requestWillBeSent for the Document fires before
Page.frameNavigated. Playwright's FrameManager needs the request
present as pendingDocument before the frame commits, otherwise
currentDocument lands with request=undefined.
Subresource events and the Document's responseReceived /
loadingFinished stay where they were.
Verified end-to-end on both clients against example.com (200),
google.com (200), httpbin/404 (404).
Three DOM-completeness gaps that broke jQuery 1.11.3 init and Playwright
actionability:
1. Element.attributes was undefined, so legacy IE feature-detect code
like div.attributes[eventName].expando threw TypeError. Added a
NamedNodeMap-like backed by a new attribute_names DOM op (length,
numeric index, string name lookup, getNamedItem/setNamedItem,
iterable).
2. Standard on* event handler properties were missing from window, so
('on' + ev) in window returned false and forced jQuery into the
legacy IE branch. Initialised the 79 HTML-spec on* handlers to null.
3. getBoundingClientRect returned the same {x:8,y:8,width:100,height:20}
literal for every element, so Playwright's hit-testing couldn't tell
elements apart. Replaced with a deterministic synthesis from _nid
mapped into a 12-column grid sized to a 1280x720 viewport. Still not
pixel-accurate (no layout engine), but stable across reads and
distinct per node.
set_default_paths only resolves OpenSSL's compile-time CA dirs, which
exist on Linux but are empty on Windows. The stealth client's cert store
ended up empty there and every TLS handshake failed with
CERTIFICATE_VERIFY_FAILED inside BoringSSL handshake.cc.
CertStore::default() uses wreq's bundled webpki-root-certs (Mozilla's
trust list), which behaves the same on every platform and removes the
dependency on a system CA store.
obscura was calling String::from_utf8_lossy on every response body, which
silently corrupts every non-UTF-8 page (GBK, Big5, Shift-JIS, Windows-125x,
EUC-KR, ISO-8859-x). Whole language families came through as replacement
characters.
Add an encoding module (obscura-net::encoding) that picks the decoder
the way real browsers do per HTML5 § 8.2.2.4:
1. HTTP Content-Type charset=... parameter (authoritative when present)
2. <meta charset> or <meta http-equiv=Content-Type> sniffed from the
first 1KB of the body (HTML only)
3. UTF-8 fallback
Use it in:
- obscura-net Response::text (signature changed: String, not Result)
- obscura-browser page.rs HTML body decode (the primary scrape path)
- obscura-browser page.rs JS body decode
- obscura-browser page.rs CSS body decode
For JS / CSS we deliberately skip the meta sniff: those bodies are not
HTML and any '<meta charset=...>' substring inside them is not a charset
declaration.
Adds 7 unit tests for the encoding module covering header precedence,
quoted charset values, meta charset sniff, http-equiv form, the 1KB
sniff window, and a round-trip decode of GBK bytes. Adds a Python
feature test exercising file:// + meta charset and a live HTTP server
serving Shift_JIS with the wrong meta tag.
Closes#113.
Three independent gaps that hurt browser-based MCP clients connecting
to obscura's CDP endpoint:
- Runtime.enable returned "No page" when called before a session was
attached. puppeteer-extra's FrameManager.initialize fires it on the
browser-level WebSocket BEFORE any page target exists, so the
standard puppeteer.connect / browser.newPage flow died on connect.
Now reply with {} when no session is attached and let the next
Target.attachToTarget emit executionContextCreated.
- Runtime.evaluate had no per-call timeout. A long synchronous JS
expression held the global V8 lock indefinitely and starved every
other CDP command on the session. Honor the CDP `timeout` parameter
and default to 30s (Chrome's protocolTimeout). Returns a clear error
on expiry so client-side deadlines fire cleanly instead of hanging.
- Target.getTargetInfo for the browser target omitted canAccessOpener.
Strict CDP clients (chromiumoxide) panic when the field is missing.
Fill it in (false for the browser target itself).
Partial fix for #122. The remaining items (networkIdle hard-deadline
on SPAs, /json/list hardcoded data, docs on cheap-vs-queued methods)
need bigger or non-code changes and are deferred.
- Element.querySelector / querySelectorAll scope to the receiver subtree
via a new query_selector_scoped Rust op, not always document root.
Same for DocumentFragment.
- setTimeout respects its delay argument via a new async op_sleep.
setInterval is a real repeating timer instead of an alias to setTimeout.
- document.forms / images / links return live querySelectorAll results
instead of hardcoded [].
- HTMLFormElement subclasses Element with .elements / .length, and form
nodes are wrapped as HTMLFormElement so instanceof works. submit()
stays on Element.prototype to preserve the existing submit-event
dispatch and form-data navigation path.
- Element.prepend has a real implementation.
- Node.isEqualNode does a structural compare (nodeType, nodeName,
nodeValue, attributes, recursive children) instead of nid identity.
- Fix long-standing insert_before arg-order bug: bootstrap.js was
passing (parent, new, ref) but _dom only forwards two args, so ref
was silently dropped. Pass (new, ref) directly. Same fix in
replaceChild.
Adds 3 selector unit tests and 6 runtime regression tests covering
the JS shim fixes.
Closes#105.
The OPTIONS preflight hardcoded Access-Control-Allow-Headers to just
Content-Type. Any browser MCP client sending mcp-protocol-version (per
the MCP spec) or Authorization / X-API-Key (typical for hosted
deployments) failed the preflight and the browser blocked the actual
request.
Add those three to the allow-list and a 24h Max-Age so browsers don't
re-preflight every call. Adds an integration test that drives the
server with a real preflight request.
Closes#175.
Playwright fails locator actions on Obscura for two avoidable reasons
that are separate from the missing layout engine. Both have trivial
fixes that unblock the actionability polling loop.
- Element.checkVisibility(opts) was undefined. Playwright >= 1.25 calls
it before every input event and retries until its action timeout if
the method isn't a function. Add a stub that returns true. Without a
layout pass we can't compute it for real, but a permissive stub still
beats the undefined-function path.
- ARIA reflection properties were missing on Element (role, ariaLabel,
ariaSelected, ariaExpanded, ariaHidden, ariaChecked, ariaDisabled,
ariaRoleDescription). Playwright's getByRole and getByLabel locators
read them. Reflect through to the underlying aria-* attributes for
both read and write.
- DOM.getBoxModel and DOM.getContentQuads ignored the objectId param.
Playwright passes only objectId (from a prior DOM.resolveNode), and
the handlers silently defaulted to node 0, so dispatchMouseEvent
always clicked the wrong element. Extract a resolve_node_id helper
that honors nodeId, backendNodeId, and objectId in that order.
- Page.captureScreenshot and Page.captureSnapshot now return a clear
"not supported, use a real Chromium for the screenshot leg" error
instead of the generic "Unknown Page method" fallback. Matches the
pattern already in place for Page.printToPDF.
Partial fix for #45. The architectural blocker (no layout engine, so
multi-element click coordinates are wrong) is unchanged.
Merges PR #165 and PR #172.
PR #165: The CDP spec marks canAccessOpener as a required boolean on
TargetInfo but Obscura was omitting it. Strict Rust clients like
chromiumoxide reject the event entirely and panic with "Created target
not present" when the field is absent. Added canAccessOpener: false
to every TargetInfo emitted by setDiscoverTargets, getTargets,
createTarget, attachToBrowserTarget, attachToTarget, and getTargetInfo.
PR #172: Added a cross-compile matrix entry for aarch64-unknown-linux-gnu
running on ubuntu-22.04 x86_64, producing obscura-aarch64-linux-cross.
This is a fallback for when GitHub ARM runners are unavailable without
removing the native ARM build.
- about: URL scheme now handled in navigate_single instead of
hitting validate_url and failing with "Forbidden URL scheme"
- DOM.getBoxModel calls getBoundingClientRect via JS instead of
returning hardcoded pixel coords
- DOM.getContentQuads added (was missing, needed by Playwright click)
- globalThis._wrap exposed so CDP eval code can look up nodes by id
- globalThis.self = globalThis for webpack/bundler worker detection
document.readyState was hardcoded to "complete" so scripts checking
readyState === 'loading' always took the wrong branch. Fixed by using a
JS global (__documentReadyState__) that Rust sets to "loading" before
any scripts execute, "interactive" before DOMContentLoaded fires, and
"complete" after load.
document.addEventListener and document.dispatchEvent on Document were
no-ops, so DOMContentLoaded listeners registered by inline scripts were
never called. Added proper listener storage and dispatch on Document.
Adds document.getElementsByName() (was missing entirely).
Page.reload is now implemented (re-navigates to the current URL),
resolving "Unknown Page method: reload" errors from Playwright.
data: URL navigation now works, fixing Playwright's goto("data:text/html,...")
pattern that previously errored with "invalid format".
Page.navigate logic extracted into do_navigate() so reload can reuse it.
Fixes#137, #146 (SPA bootstrap lifecycle ordering).
GHSA-8v6v-g4rh-jmcm: op_fetch_url validated the initial URL through
validate_fetch_url, but reqwest auto-followed up to 10 redirects with
no further checks. An attacker-allowed origin could 302 to
http://127.0.0.1 (or any private/internal address) and read the body
back to page JS. Real-world impact: AWS instance metadata, kube
debug consoles, internal admin endpoints with permissive CORS all
become readable from a page hosted anywhere on the open web.
Disable reqwest's auto-redirect and follow manually. Apply
validate_fetch_url to every redirect target before requesting it.
Cap the chain at 10 hops to match prior behaviour. Browser-standard
method/body rules apply on cross-method redirects (301/302/303
downgrade to GET with no body, 307/308 preserve).
Verified vs unpatched build (pr69):
- Patched: attacker-page -> 302 -> http://127.0.0.1:N is blocked
with "Redirect to forbidden URL blocked: Access to private IP..."
before the internal hop runs. Legitimate 3-hop redirect chain
between allowed origins still completes.
- Unpatched: internal-service body leaks through to page JS via
fetch().then(r => r.text()).
GHSA-q55h-vfv9-qcr5 patched Page.navigate to refuse file:// URLs
unless --allow-file-access was set, but Target.createTarget accepted
the same scheme and called page.navigate(url) directly, bypassing
the policy. A CDP client could call Target.createTarget {url: file://...}
then Runtime.evaluate to read document.body.textContent and recover
arbitrary local files.
Apply the same gate at the Target.createTarget entrypoint. Pull
the file-scheme detector into a shared util module so both domains
stay in lock-step instead of one drifting from the other.
Verified blocked vs the unpatched build (pr69):
- Patched: Target.createTarget {url: file:///tmp/secret} errors
"Target.createTarget to file:// is disabled. ..." before any
navigation runs.
- Unpatched: same call returns the file contents via
Runtime.evaluate {expression: "document.body.textContent"}.
Two improvements while investigating #147:
1. `obscura --version` / `-V` now print the CARGO_PKG_VERSION instead of
erroring out. The issue reporter explicitly called out that there was
no version flag, and the existing `obscura serve` banner is the only
way to discover the version.
2. New regression test in obscura-js verifies that a TypeError thrown
from one script does not poison the runtime: subsequent scripts and
DOM queries must continue to work. The reporter's symptom was that
offside.js crashing collapsed --dump text output to 1 byte; that
cascade should never happen and this test pins it down.
The classList error itself is offside.js's own bug (it indexes into the
querySelectorAll result without checking for an empty array) and only
fires when the .slideout-navigation element is absent from the DOM.
On current main that element parses correctly and offside.js runs
without error against a captured copy of the affected page.
The previous workaround in page.rs rewrote socks5:// to http:// before
handing the proxy to the stealth client. That works for Clash-style
mixed-mode proxies, but breaks plain SOCKS5 servers like `ssh -ND`,
producing a confusing `TunnelUnexpectedEof` deep in the network layer.
The stealth client is backed by wreq, which does not speak SOCKS5,
so silently substituting schemes can never be correct in general.
Refuse the combination at CLI startup with a clear message that tells
the user what to do (use --proxy http://... or drop --stealth) and
revert the rewrite in Page::new. SOCKS5 without --stealth still works
through the reqwest path.
Uses GitHub's native ubuntu-22.04-arm runner so it's a native build
matching the same glibc 2.35 floor as the x86_64 target. README gets
the matching install snippet.
GHSA-q55h-vfv9-qcr5. Any process able to reach the CDP port could
send Page.navigate {"url":"file:///etc/shadow"} and read arbitrary
local files (the process's UID's worth of read access). When obscura
runs as root inside Docker or as a systemd service that translates
straight into /etc/shadow, /root/.ssh/id_rsa, /proc/1/environ, etc.
Validate at the CDP boundary, not in obscura-net, so direct CLI
callers ("obscura fetch file://...") keep working. A new
BrowserContext.allow_file_access flag defaults to false; Page.navigate
returns an error when the requested URL is file:// and the flag is
off. CdpContext::new_with_security threads the flag from the server
entry point, and "obscura serve --allow-file-access" opts in for
local-HTML testing workflows.
Tested:
- Default obscura serve: CDP Page.navigate {file:///etc/passwd} ->
refused with descriptive error.
- obscura serve --allow-file-access: same call succeeds.
- obscura fetch file://... CLI path: unchanged, still reads the file.
- All earlier PoCs (GHSA-6pj9 JS->file SOP, GHSA-4789 sub-resource
SSRF, GHSA-4789 CSS template injection) still pass.
- cargo test --workspace clean.
Three findings from GHSA-4789-rq5x-jv3c.
HIGH-1: Sub-resource fetches (<script src>, <link rel=stylesheet>) had
no scheme gate, so an http(s) page could include
<script src="file:///etc/passwd"> and the body would come back as
script source. Added subresource_allowed() that mirrors the top-level
file:// SOP rule: file: sub-resources only load when the page itself
is on file:. Off-origin schemes (data:, javascript:, etc.) are
refused regardless. Private/loopback hosts were already blocked at
the obscura-net layer; this closes the file:// hole on top of that.
HIGH-2: CSS pulled from the network was string-formatted into a JS
template literal with a three-rule escaper (backslash, backtick,
${). That escaper did not handle U+2028 / U+2029 which terminate
ECMAScript template literals, letting attacker CSS break out and
run arbitrary code in the page's V8 realm. Replaced the inline
escape with escape_for_js_template_literal() which also handles
U+2028, U+2029, null, carriage return, and every other C0 control.
LOW-2: navigate_with_wait_post() capped JS-driven navigation chains
at ten but returned Ok(()) when the cap was hit, so callers could
not tell a successful load from a redirect storm. Now returns
PageError::TooManyRedirects(REDIRECT_LIMIT).
Verified with three PoCs (sub-resource SSRF, CSS template literal
break with U+2028, and the original GHSA-6pj9-j389-226p reproducer)
plus negative tests (file:// direct read, http -> http JS chain) all
pass. cargo test --workspace clean.
A page on http(s) could set window.onload, assign location.href to a
file:// URL, and then read document.body from the local file because
two things lined up: obscura-net allows file: URLs, and the JS
runtime was reused across navigations so window state from the prior
origin survived into the new document's load event.
Two fixes:
1. Refuse a JS-initiated chain step when it crosses from a non-file
scheme into file:. Direct file:// navigations from the CLI/CDP
still work because the gate only fires on the take_pending_nav
path. file -> file chains also still work.
2. Drop and rebuild the V8 runtime on every navigation instead of
only rebinding globalThis.document. Even within the same origin
we should not carry over window.onload, custom properties, or
registered event listeners from the previous document.
Tested with the PoC from the report: pages on http://127.0.0.1.nip.io
no longer leak the contents of file:///tmp/.../pwn.txt.
Adds a workflow that triggers on v* tag pushes and:
- Logs into Docker Hub using DOCKERHUB_USERNAME + DOCKERHUB_TOKEN secrets
- Builds the image from the existing Dockerfile
- Tags it with both the version (e.g. 0.1.4) and 'latest'
- Pushes both tags
GitHub Actions cache is used for layers to speed up subsequent builds.
Adds a multi-stage Dockerfile (rust:1-slim-bookworm builder → distroless/cc
runtime) plus a --host flag for obscura serve that defaults to 127.0.0.1
for safety but can be set to 0.0.0.0 so the container's port is reachable
via docker run -p mapping.
Dockerfile cherry-picked from #84. The SPA chunk loading commit from that
PR was left out — verified it didn't actually fix the Next.js webpack
errors on kiwi.com (#86).
SGavrl