shuttle
AGENTS.md43 starsChanged 33 days agoBSD-3-ClauseView on GitHub
Shuttle (Shuttle Browser, by Silkweave) is a deep fork of Chromium for AI agents. This directory is the fork (github.com/shuttle-browser/shuttle), checked out as src/ of a gclient checkout whose root is the parent directory (.gclient, logs/, release/). gclient requires the name src. Read, in order: Run tooling from tools/shuttle/ (scripts/...). The fork is public: run /dev:gatekeeper + scoped gitleaks before pushing; only our own additions are scanned, never Chromium's tree or history. Real path for builds. Siso keys its state…
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,542 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repositoryOne other file it gives its agents.
- style-guideSkill
The file
We split this AGENTS.md into the rules it gives an agent
3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 13
Run tooling from tools/shuttle/ (scripts/...).
- Very specific: line 23
check: Mac gn gen of the dev out dir ($SHUTTLE_BUILD_DIR, default out/Dev) passes (from the real path).
- Distinctive: line 26
test: skip (unit/browser test targets are pruned)
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Shuttle (fork root)
Shuttle (Shuttle Browser, by Silkweave) is a deep fork of Chromium for AI agents. This directory is the fork
(github.com/shuttle-browser/shuttle), checked out as src/ of a gclient checkout whose root is the parent directory
(.gclient, logs/, release/). gclient requires the name src.
Read, in order:
_docs/AGENTS-TEAM.md: team, status, rules (private, git-ignored, only on the lead's machine)_docs/NEXT-SESSION.md: your handofftools/shuttle/AGENTS.md: tooling, build, removal workflow, gotchas
Run tooling from tools/shuttle/ (scripts/...). The fork is public: run /dev:gatekeeper + scoped gitleaks before pushing;
only our own additions are scanned, never Chromium's tree or history.
Real path for builds. Siso keys its state by absolute path. If this session's cwd is a symlink (e.g.
…/shuttle/src), cd to the real checkout (pwd -P) before any raw gn / siso / autoninja command; a build seen
under a second path drops the build cache (full rebuild, about a day). The tools/shuttle/scripts/* build scripts already
resolve the real path.
Wrapup Config
- check: Mac
gn genof the dev out dir ($SHUTTLE_BUILD_DIR, defaultout/Dev) passes (from the real path); if fork browser code changed since the last build, a green dev build and the CDP smoke test; Linux state confirmed with the Linux session (green or open errors listed) - test: skip (unit/browser test targets are pruned)
- security: the fork is public:
/dev:gatekeeper(gatekeeper --auditfrom this directory, our additions only via.claude/gatekeeper.json) +gitleaks git --staged/--log-opts=github/main..HEAD, alwaysGOMAXPROCS=2 nice -n 19 taskpolicy -c background, never Chromium's full tree/history - push: work on a branch (
sh/<topic>), push the branch togithuband open a PR tomain; the maintainer reviews and merges every PR (Toby, 2026-10-17). Tooling commits unnumberedshuttle/: …, browser changes = numbered fork commits; never commitbuild/util/MINIBRANCH - version_bump: no
- publish: no
- docs: public
tools/shuttle/docs/(TOOLING gotchas, REMOVAL-LOG per fork commit, LINUX, plans) +tools/shuttle/AGENTS.md; private_docs/AGENTS-TEAM.mdStatus - next_session: always rewrite
_docs/NEXT-SESSION.md(Linux session: its own file in_docs/): goal, read first, current state, ground rules, steps, done-when, open questions; running builds/shells listed last - team: tell the Linux session what you committed/pushed; add the session's resets and findings to
_docs/AGENTS-FLOW.md - frontend_smoke: no (CDP smoke test is part of check)
- co_authored_by: no (global)
- changelog: no
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

