Add mention-in-issue and mention-in-pr variants with sandbox disabled#367
Add mention-in-issue and mention-in-pr variants with sandbox disabled#367
Conversation
… variants Co-authored-by: strawgate <6384545+strawgate@users.noreply.github.com>
📝 WalkthroughWalkthroughThis PR adds two new GitHub Actions workflow variants—"mention-in-issue-no-sandbox" and "mention-in-pr-no-sandbox"—that operate without sandbox restrictions, enabling direct Docker access. Includes generated lock files, workflow markdown definitions, documentation, example configurations, navigation updates, and script exclusions. Changes
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Possibly related PRs
Suggested reviewers
Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches
🧪 Generate unit tests (beta)
Comment |
There was a problem hiding this comment.
🧹 Nitpick comments (1)
.github/workflows/gh-aw-mention-in-pr-no-sandbox.md (1)
93-97: Note: eval usage with dynamic setup commands.Using
evalwith theSETUP_COMMANDSinput is a common CI/CD pattern for flexible setup but carries inherent risk. The risk is mitigated here by:
- Workflow trigger restricted to admin/maintainer/write roles
- setup-commands is a workflow input, not arbitrary user input
- Pattern is standard for dynamic CI setup
This is acceptable in context but worth noting for security awareness.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-mention-in-pr-no-sandbox.md around lines 93 - 97, The workflow uses eval "$SETUP_COMMANDS" (env SETUP_COMMANDS) which can be dangerous; replace the eval usage in the "Repo-specific setup" step by validating or sanitizing SETUP_COMMANDS and executing it via a safer mechanism (e.g., run with bash -lc or a controlled wrapper) and/or implement a whitelist or explicit parsing of allowed subcommands before executing; update the step that references SETUP_COMMANDS and the run: eval "$SETUP_COMMANDS" invocation to call the safer executor and add input validation logic to mitigate injection risk.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Nitpick comments:
In @.github/workflows/gh-aw-mention-in-pr-no-sandbox.md:
- Around line 93-97: The workflow uses eval "$SETUP_COMMANDS" (env
SETUP_COMMANDS) which can be dangerous; replace the eval usage in the
"Repo-specific setup" step by validating or sanitizing SETUP_COMMANDS and
executing it via a safer mechanism (e.g., run with bash -lc or a controlled
wrapper) and/or implement a whitelist or explicit parsing of allowed subcommands
before executing; update the step that references SETUP_COMMANDS and the run:
eval "$SETUP_COMMANDS" invocation to call the safer executor and add input
validation logic to mitigate injection risk.
Some projects need Docker access during agent runs (for example in
setup-commands), which is blocked by the default sandboxed mention workflows. This PR adds no-sandbox variants without replacing the existing sandboxed workflows.What changed
gh-aw-mention-in-issue-no-sandboxgh-aw-mention-in-pr-no-sandboxsandbox.agent: false, so the agent runs directly on the runner and can access Docker.gh-agent-workflows/mention-in-issue-no-sandbox/andgh-agent-workflows/mention-in-pr-no-sandbox/..github/workflows/for both no-sandbox variants.mkdocs.ymlnavigation to include both no-sandbox workflow docs.scripts/dogfood.shEXCLUDED_WORKFLOWSto exclude both no-sandbox variants and avoid duplicate/aitrigger conflicts in this repository.Consumer usage
Warning
Disabling the agent sandbox removes the sandbox firewall and filesystem isolation. Use these variants only when Docker access is required.