Surface plugins in GitHub Actions and console run summaries

[Copilot]

ExperimentConfiguration.plugins was populated by PR #753 but never rendered in run summaries — reviewers had no way to see which plugins were loaded from the job summary or console output.

Changes

src/bcbench/results/display.py

• Added - Plugins: line to create_github_job_summary header, placed after - Custom Agent:, treating plugins as a list exactly like mcp_servers (, -joined, falls back to None)
• Added matching Plugins: line to create_console_summary for parity

f"- Plugins: {', '.join(results[0].experiment.plugins) if results[0].experiment and results[0].experiment.plugins else 'None'}",

tests/test_result_hierarchy.py

• Fixed test_github_summary_renders_markdown: split the "- Custom Agent: N/A\n\n## Result Summary" adjacency assertion since - Plugins: is now the last header line; also asserts - Plugins: None when no plugins are set
• Added test_github_summary_shows_plugins_when_present: verifies both plugin entries ("<name>@<sha>" / "<name>@local") appear joined in the rendered content

[Copilot]

Pull request overview

Surfaces configured plugins in console and GitHub Actions run summaries.

Changes:

• Adds plugin lists to both summary formats.
• Adds GitHub summary assertions for configured and absent plugins.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
src/bcbench/results/display.py	Renders plugin configuration in summaries.
tests/test_result_hierarchy.py	Tests GitHub plugin summary output.

[Copilot]

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated no new comments.

Suppressed comments (2)

src/bcbench/results/display.py:34

• PluginConfig.name is unrestricted, so a valid plugin record may contain Rich markup such as [red]. Interpolating it inside a markup-enabled string makes Rich interpret the name; unknown or mismatched tags can raise MarkupError and prevent the console summary from being emitted, while valid tags alter the displayed identifier. Render the dynamic value with markup disabled (or escape it) instead.

    console.print(f"Plugins: [bold]{', '.join(results[0].experiment.plugins) if results[0].experiment and results[0].experiment.plugins else 'None'}[/bold]")

src/bcbench/results/display.py:102

• Plugin names are not constrained to Markdown-safe single-line text. For example, a valid local plugin name containing a newline ends this list item and creates arbitrary additional Markdown, so the job summary no longer shows the recorded plugin identifier faithfully. Escape Markdown/HTML metacharacters and normalize line breaks before adding plugin records to the summary.

            f"- Plugins: {', '.join(results[0].experiment.plugins) if results[0].experiment and results[0].experiment.plugins else 'None'}",