Repository navigation
Start sessions automatically for new development clones - #705
Conversation
Treat completed allowed-owner clones under the configured roots as session candidates. Keep the interactive marker requirement for linked worktrees and continue excluding internal agent worktrees. Adopt existing unmarked clones on upgrade so the change starts sessions only for clones that appear afterward.
A clone seen before its origin was configured was permanently ignored. Wait for the checkout to finish before recording whether its owner is allowed, and cover the incomplete-to-finished transition with a real Git regression.
|
@codex review |
|
@codex security review |
|
@coderabbitai review |
|
@greptileai review |
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
Security findingsAdvisory findings (1)ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
|
Important
This repository does not receive automatic reviews because it has fewer than 10 stars.
 |
|
@coderabbitai review |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
🛡️ Codex Security Review
Here are some automated security review suggestions for this pull request.
Reviewed commit: 2582056c09
ℹ️ About Codex security reviews in GitHub
This is an experimental Codex feature. Security reviews are triggered when:
- You comment "@codex security review"
- A regular code review gets triggered (for example, "@codex review" or when a PR is opened), and you’re opted in so security review runs alongside code review
Once complete, Codex will leave suggestions, or a comment if no findings are found.
Keep automatic clone sessions on dev boxes while requiring the interactive marker for both agents on both admin jumpboxes. Fence the watcher before changing its policy and code, and publish policy files atomically so a failed update cannot enable unapproved startup hooks. Cover both admin installation paths and failure ordering without restarting interactive sessions.
ejc3
changed the title
|
@codex review |
|
@codex security review |
|
@coderabbitai review |
|
@greptileai review |
|
Codex Review: Didn't find any major issues. What shall we delve into next? Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
🛡️ Codex Security ReviewSecurity review completed. No security issues were found in this pull request. Reviewed commit: Only the user who started this review can view the report in Codex. ℹ️ About Codex security reviews in GitHubThis is an experimental Codex feature. Security reviews are triggered when:
Once complete, Codex will leave suggestions, or a comment if no findings are found. |
|
|
|
|
Merged and deployed as 3ecf74b. Terraform applied the six reviewed script updates and two administrator convergence triggers; the full follow-up plan is empty and /infra/applied-status is clean. All seven active watchers now run the merged code and have completed the upgrade baseline. Both jumpboxes enforce the marker in root-owned policy and actual process arguments. All 12 baseline administrator Claude/Codex processes survived; ARM and container watchers kept their PIDs. Existing clones were seeded without launching a burst of sessions. Colton is live on the new behavior. The stopped x86 host was left stopped, with its startup script updated for the next boot. The running ARM watcher was updated with an atomic copy of only the exact merged script, without package commands or a full bootstrap. |
Newly cloned repositories on the development servers now get Claude and Codex sessions automatically once the checkout is complete and its owner is allowed. Linked worktrees still require the interactive marker, and internal agent worktrees remain excluded.
Both administrator jumpboxes retain the explicit marker requirement for both agents, including cold bootstrap. An allowed repository origin does not approve running its startup hooks with administrator credentials. When the admin policy changes, installers stop only the watcher before publishing the guarded policy and new code; policy writes are atomic, and failures stop installation. Existing interactive sessions remain independent of the watcher.
The upgrade records existing clones without opening sessions for all old checkouts. Unfinished clones are retried before recording an owner rejection, so a clone seen before its origin is written is detected when it finishes.
Validation: 69 focused ARM tests pass, plus Terraform formatting and helper shell syntax checks. Independent high code and security reviews passed. The clone-race, upgrade-baseline, admin-marker, and installer-ordering tests were observed failing when the corresponding guards were removed. Final CI passed all 69 tests, and Codex code and security reviews found no issues on the published head. CodeRabbit and Greptile provided no usable final-head review and were not counted as approvals.