fix: update assign_to_agent tests to use handler factory pattern

[Copilot]

Fixes the failing js CI job (job ID 70179864906, run 24061973971) on the copilot/add-cli-proxy-feature-flag branch.

Root Cause

The main() function in assign_to_agent.cjs was refactored from a standalone item processor to a handler factory pattern (returning a message handler function for the safe output handler manager), but the test suite (assign_to_agent.test.cjs) still used the old standalone interface with env vars and direct main() calls.

Changes

• assign_to_agent.test.cjs: Added a STANDALONE_RUNNER code snippet that simulates what the safe output handler manager does — builds handler config from env vars, calls main(config) as a factory to get the handler function, processes assign_to_agent items through the handler, writes summary, sets step outputs, and calls core.setFailed() on errors
• Replaced all await main() eval calls (44 occurrences) with ${STANDALONE_RUNNER} so every test exercises the production factory interface
• Fixed assertion: updated "encounters errors" to "encounters auth errors" to match the actual production log message

No changes were made to assign_to_agent.cjs production code.

Testing

• All 41 previously-failing tests now pass (3 intentionally skipped)
• Full JS test suite: 6364/6365 pass (1 pre-existing network test failure unrelated to these changes)

[pelikhan]

@copilot fix the test, remove the special handling added to the product code.

[Copilot]

@copilot fix the test, remove the special handling added to the product code.

Done in 2728677. Reverted the backward-compat code from assign_to_agent.cjs (production code is unchanged). Updated assign_to_agent.test.cjs with a STANDALONE_RUNNER snippet that simulates the handler manager flow (build config from env vars → call main(config) as factory → process items → write summary → set outputs). Also fixed the "encounters auth errors" assertion to match the actual production log message.

Warning

Firewall rules blocked me from connecting to one or more addresses (expand for details)

I tried to connect to the following addresses, but was blocked by firewall rules:

• https://api.github.com/repos/github/gh-aw/contents/.github%2Fworkflows%2Faudit-workflows.md
  • Triggering command: /opt/hostedtoolcache/node/24.14.1/x64/bin/node /opt/hostedtoolcache/node/24.14.1/x64/bin/node --experimental-import-meta-resolve --require /home/REDACTED/work/gh-aw/gh-aw/actions/setup/js/node_modules/vitest/suppress-warnings.cjs --conditions node --conditions development /home/REDACTED/work/gh-aw/gh-aw/actions/setup/js/node_modules/vitest/dist/workers/forks.js . it git rev-�� --reverse origin/fallback-branch..HEAD 1/x64/bin/git -m Initial commit de/node/bin/git git (http block)
• invalid.example.invalid
  • Triggering command: /usr/lib/git-core/git-remote-https /usr/lib/git-core/git-remote-https origin https://invalid.example.invalid/nonexistent-repo.git git comm�� -m lure test commit e_modules/.bin/git ranch main bin/git git conf�� user.email test@test.com /git -b feature-branch 1/x64/bin/git git (dns block)

If you need me to access, download, or install something from one of these locations, you can either:

• Configure Actions setup steps to set up my environment, which run before the firewall is enabled
• Add the appropriate URLs or hosts to the custom allowlist in this repository's Copilot coding agent settings (admins only)

[Copilot]

Pull request overview

Updates the assign_to_agent JavaScript tests to exercise the refactored handler-factory interface (main(config) -> handler) instead of calling main() as a standalone processor, fixing failing CI on the branch.

Changes:

• Added an in-test “standalone runner” that builds a handler config, calls main(config) to get a message handler, then processes assign_to_agent items from GH_AW_AGENT_OUTPUT.
• Replaced prior await main() eval calls across the suite to run through the factory-based flow.
• Updated an assertion to match the production log message (“encounters auth errors”).

Show a summary per file

File	Description
actions/setup/js/assign_to_agent.test.cjs	Refactors tests to run assign_to_agent.cjs via the handler-factory pattern using a shared runner snippet.

Copilot's findings

Tip

Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

• Files reviewed: 1/1 changed files
• Comments generated: 2

[Copilot]

The test runner re-implements agent output loading with raw fs.readFileSync + JSON.parse and assumes .items exists. Production uses loadAgentOutput() (actions/setup/js/load_agent_output.cjs) which handles missing files, empty content, JSON parse errors, and validates the items array. Consider using that shared helper in STANDALONE_RUNNER (or at least matching its validation/error handling) so the tests don’t drift from production behavior.

Suggested change

	const _fs = require("fs");
	const _agentOutput = JSON.parse(_fs.readFileSync(_agentOutputFile, "utf8"));
	const _items = _agentOutput.items.filter(i => i.type === "assign_to_agent");
	const { loadAgentOutput } = require("./load_agent_output.cjs");
	const _agentOutput = loadAgentOutput(_agentOutputFile);
	const _items = Array.isArray(_agentOutput?.items)
	? _agentOutput.items.filter(i => i.type === "assign_to_agent")
	: [];

[Copilot]

If STANDALONE_RUNNER is meant to simulate the safe output handler manager, the output keys should match the production manager’s exported outputs (assign_to_agent_assigned, assign_to_agent_assignment_errors, assign_to_agent_assignment_error_count in actions/setup/js/safe_output_handler_manager.cjs). Right now the runner uses legacy keys (assigned, assignment_errors, assignment_error_count), which can make the tests pass while diverging from the real workflow interface.