Fix Playwright MCP TOML sandbox configuration

[Copilot]

Playwright MCP configurations rendered as TOML omitted Chromium’s --no-sandbox flag, unlike JSON-rendered configurations. This could prevent browser startup in sandboxed environments.

• TOML renderer

  • Adds --no-sandbox to Playwright MCP entrypoint arguments.

• Regression coverage

  • Verifies TOML output includes the required Docker security flags and browser launch flag.

entrypointArgs = [
  "--output-dir",
  "/tmp/gh-aw/mcp-logs/playwright",
  "--no-sandbox"
]

---

Warning

Firewall blocked 1 domain

The following domain was blocked by the firewall during workflow execution:

• github.com

To allow these domains, add them to the network.allowed list in your workflow frontmatter:

network:
  allowed:
    - defaults
    - "github.com"

See 网络 Configuration for more information.

Generated by 👨‍🍳 PR Sous Chef · pi · gpt54 · 33.5 AIC · ⌖ 8.65 AIC · ⊞ 8.8K · ◷
_{Comment /souschef to run again}

[pelikhan]

@copilot update an agentic workflow that uses pnaywrigjt to use codex to test this

[Copilot]

@copilot update an agentic workflow that uses pnaywrigjt to use codex to test this

Updated smoke-codex to use Playwright MCP and invoke browser_navigate/browser_snapshot; its regenerated Codex TOML configuration now exercises --no-sandbox. Included in 29a354e.

[Copilot]

🟡 Changes recommended

Custom Playwright arguments now produce invalid TOML due to duplicate comma emission.

Once you've addressed the issues Copilot identified, you can request another Copilot review.

Pull request overview

Adds Chromium’s --no-sandbox flag to TOML-rendered Playwright MCP configurations and exercises it through the Codex smoke workflow.

Changes:

• Adds the Playwright launch flag and regression coverage.
• Migrates the Codex smoke test from Playwright CLI to MCP.
• Regenerates the compiled workflow.

File summaries

File	Description
pkg/workflow/mcp_renderer_builtin.go	Adds --no-sandbox to TOML output.
pkg/workflow/mcp_config_refactor_test.go	Checks required Playwright flags.
.github/workflows/smoke-codex.md	Tests Playwright through MCP.
.github/workflows/smoke-codex.lock.yml	Regenerates the compiled workflow.

Review details

• Files reviewed: 4/4 changed files
• Comments generated: 1
• Review effort level: Balanced

---

💡 Add a code-review agent skill for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Fixed in 2bc3036. The built-in --no-sandbox argument is now unterminated so writeArgsToYAML can insert the separator before custom arguments; regression coverage verifies the resulting sequence.

[gh-aw-bot]

@copilot Please refresh this PR for forward progress. Re-run the pr-finisher skill, address any remaining review feedback, and update the branch if needed.

Generated by 👨‍🍳 PR Sous Chef · pi · gpt54 · 11 AIC · ⌖ 8.68 AIC · ⊞ 8.8K · ◷
_{Comment /souschef to run again}

[gh-aw-bot]

@copilot There is still review feedback on this PR about the TOML rendering regression. Please address the duplicate-comma TOML issue noted by Copilot review, refresh the branch if needed, and then run the pr-finisher skill before handing back.

Current review feedback, newest first:

• Copilot review: custom Playwright arguments now produce invalid TOML due to duplicate comma emission — Fix Playwright MCP TOML sandbox configuration #56800

Run: https://github.com/github/gh-aw/actions/runs/33246040923

Warning

Firewall blocked 1 domain

The following domain was blocked by the firewall during workflow execution:

• github.com

To allow these domains, add them to the network.allowed list in your workflow frontmatter:

network:
  allowed:
    - defaults
    - "github.com"

See 网络 Configuration for more information.

Generated by 👨‍🍳 PR Sous Chef · pi · gpt54 · 33.5 AIC · ⌖ 8.65 AIC · ⊞ 8.8K · ◷
_{Comment /souschef to run again}

[Copilot]

@copilot There is still review feedback on this PR about the TOML rendering regression. Please address the duplicate-comma TOML issue noted by Copilot review, refresh the branch if needed, and then ru...

The duplicate-comma regression is already fixed in 2bc3036, with focused custom-argument coverage. I refreshed the branch with main (108a138) and re-ran formatting, lint, unit tests, full tests, and workflow recompilation successfully.