fix: pre-landing review fixes

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
This commit is contained in:
garrytanandcapy-ai[bot] committed 2026-10-26 02:03:09 +00:00
1 parent 6d0d2691d5
commit e965b6cc5f
2 files changed
+8 -6

No files matched your search

+7 -5
View File
@@ -1327,8 +1327,9 @@ the latest release 1.64.1 (Chromium 156) and the 1.65.0 alpha of 2026-10-24
fails it with `ERR_CERT_REVOKED` once its CRLSet arrives). gstack will move to
the first Playwright release whose Chromium carries the removal. Until then,
don't do signed-in work in this engine on a network you do not trust (sessions
imported earlier persist in the headed profile, the `BROWSE_EXTENSIONS_DIR`
profile, `state save` files and, with `BROWSE_PERSIST_STATE=1`, the headless
imported earlier persist in each project's headed profile (`$B profiles list`
shows them; new projects copy logins from the old shared
`~/.gstack/chromium-profile`), the `BROWSE_EXTENSIONS_DIR` profile, `state save` files and, with `BROWSE_PERSIST_STATE=1`, the headless
session state), and prefer Aside for signed-in work where you have it.
To check a build, run `$B stop`, then the same `goto` in the default headless
@@ -1336,9 +1337,10 @@ mode with `BROWSE_EXTENSIONS_DIR` and `BROWSE_PROXY_URL` unset, the proxy
variables `HTTPS_PROXY`, `HTTP_PROXY` and `ALL_PROXY` unset in both cases
(Chromium reads those itself, even when gstack sets no proxy), and
`GSTACK_CHROMIUM_PATH` unset (the bundled build) or set to the build you want
to check. First confirm that a known-good HTTPS site such as
`$B goto https://example.com/` loads in the same session, because a captive
portal or an intercepting proxy fails every site with the same error. Then only
to check. Run it on a network you trust not to inspect TLS, because a captive
portal or an intercepting proxy can fail some or all sites with the same error,
and first confirm that a known-good HTTPS site such as
`$B goto https://example.com/` loads in the same session. Then only
`ERR_CERT_AUTHORITY_INVALID` on the test host means the root store no longer
trusts them. `ERR_CERT_REVOKED` only shows that the session has a CRLSet, and a
DNS, network or other error proves nothing.
+1 -1
View File
@@ -5,7 +5,7 @@
**BROWSER.md now says that gstack's own browser still trusts the compromised eMudhra intermediates, and how to check (#3338).** Chrome blocked the eight eMudhra intermediate CAs with a CRLSet push on 2026-10-17, the day their key compromise was reported, and scheduled the removal of their root, emSign ECC Root CA - G3, from the Chrome Root Store for 2026-10-27. gstack's fallback browser gets neither in time: its default headless browser never receives CRLSet updates, headed mode and `handoff` received them only about a minute after each launch in testing on Linux (#3338, from macOS, reports that gstack's profile never gets one), and the root store compiled into the bundled Chromium still trusts the root. On 2026-10-24, `$B goto https://dv-g3b-emudhra.we-have-a-problem.com/` returned 200 and "That's bad!" on gstack's pinned Playwright 1.62.1 (Chromium 151.0.7922.34). The same check against Playwright 1.64.1 (Chromium 156.0.8078.4) and the 1.65.0 alpha of 2026-10-24 (Chromium 157.0.8103.15) also loaded the page. Chrome for Testing builds from 158.0.8115.0 fail it with `ERR_CERT_AUTHORITY_INVALID`, so a Playwright bump has to wait for a release that bundles one of those builds.
### Changed
- **`BROWSER.md`**: a new "TLS trust: the bundled Chromium's root store" section, at the end of the security stack, says the fallback engine checks certificates against that compiled root store, that the headless browser never gets CRLSet updates while headed mode and `handoff` got them about a minute after launch in testing on Linux, and that the bundled build does not enforce Certificate Transparency. It names the commands that put real sessions and credentials into the browser (among them `cookie-import-browser`, `cookie-import`, `cookie`, `header`, `state load`, `storage set`, `fill`/`type`, `upload`, `handoff`, and a paired agent using `fill`/`type`), lists where imported sessions persist, says how `BROWSE_EXTENSIONS_DIR` and `GSTACK_CHROMIUM_PATH` change the picture (`handoff` always uses the bundled build, and a current Chrome only helps if it drops the root), and gives the `$B goto` check: after `$B stop`, in headless mode with no proxy variables set, and once a known-good HTTPS site loads in the same session, only `ERR_CERT_AUTHORITY_INVALID` on the test host means a build is fixed. The table of contents, the fallback-engine introduction and the `cookie-import-browser` and `handoff` command rows link to it. Aside is unaffected, because it is your own browser with its own updates.
- **`BROWSER.md`**: a new "TLS trust: the bundled Chromium's root store" section, at the end of the security stack, says the fallback engine checks certificates against that compiled root store, that the headless browser never gets CRLSet updates while headed mode and `handoff` got them about a minute after launch in testing on Linux, and that the bundled build does not enforce Certificate Transparency. It names the commands that put real sessions and credentials into the browser (among them `cookie-import-browser`, `cookie-import`, `cookie`, `header`, `state load`, `storage set`, `fill`/`type`, `upload`, `handoff`, and a paired agent using `fill`/`type`), lists where imported sessions persist, says how `BROWSE_EXTENSIONS_DIR` and `GSTACK_CHROMIUM_PATH` change the picture (`handoff` always uses the bundled build, and a current Chrome only helps if it drops the root), and gives the `$B goto` check: after `$B stop`, in headless mode with no proxy variables set, on a network that does not inspect TLS, and once a known-good HTTPS site loads in the same session, only `ERR_CERT_AUTHORITY_INVALID` on the test host means a build is fixed. The table of contents, the fallback-engine introduction and the `cookie-import-browser` and `handoff` command rows link to it. Aside is unaffected, because it is your own browser with its own updates.
### For contributors
- The edits above the cookie workflow excerpt change no line count, so the `setup-browser-cookies/SKILL.md workflow` judge input is byte-identical (prompt SHA-256 `c851e635…`, 12,805 bytes, before and after).