feat: add display_report option to disable step summary

[ashwin-ant]

Add a display_report input parameter (default: "true") that controls whether the Claude Code Report is written to the GitHub Step Summary. Setting it to "false" allows users with custom formatting solutions to avoid duplicate output.

Closes #206

Generated with Claude Code

[claude]

Nit: The description says Set to 'false' to disable — since GitHub Actions inputs are always strings, consider mentioning that any value other than "false" (including omitting the input) will keep the report enabled, since the code checks !== "false". This is fine behavior, but users might try display_report: false (boolean YAML) which GitHub Actions coerces to the string "false", so it works — but display_report: no would not disable it, which could be surprising.

A small documentation clarification in the description (e.g., "Set to the string 'false' to disable") or switching the code to check against === "true" (matching the pattern used by use_commit_signing, track_progress, etc.) would make the behavior more predictable. This is low priority since the default is "true" and "false" is the natural input.

[claude]

Review Summary

Clean, well-scoped PR. The implementation is correct and the diff is minimal. A few observations:

• The env var naming (DISPLAY_REPORT without INPUT_ prefix) is consistent with other prepare-phase env vars consumed directly by run.ts (e.g., TRIGGER_PHRASE, TRACK_PROGRESS, USE_STICKY_COMMENT).
• The !== false comparison pattern matches existing usage in update-comment-link.ts.
• No existing tests cover writeStepSummary, so the lack of new tests here is not a regression, but this area would benefit from test coverage in general.

[dsarno]

Confirmed works, thanks guys!

BEFORE: https://github.com/dsarno/unity-mcp/actions/runs/22023978435
AFTER: https://github.com/dsarno/unity-mcp/actions/runs/22046246096

FYI, those runs are on my fork of the MCP for Unity, of which I'm the lead contributor. We just hit 6K stars. I use the CI workflows to spin up an instance of Unity + python mcp, connect them, and then have Claude Code hurl a bunch of tool requests at the mcp server (stuff like setting up gameobjects and adding scripts to them). Then claude decides if the tests pass and we get that nice summary. Been doing that for many months, super helpful workflow.

And I'm sure you get asked this all the time, but if you have a program for widely used open source repos to apply for token grants, that'd be awesome. We have tens of thousands of users all over the world, and are the #1 mcp in the gaming world -- lots of people using us to learn -- free -- how to make games with the help of agents. If I can submit a proposal that'd be awesome!