fix(web_server): reap finished action subprocesses to prevent zombie accumulation

[liuhao1024]

What does this PR do?

Reaps finished dashboard action subprocesses in get_action_status() to prevent zombie process accumulation. Previously, proc.poll() was called to check liveness but proc.wait() was never issued, leaving every completed action as a <defunct> zombie for the lifetime of the web server.

Related Issue

Fixes #38032

Type of Change

• 🐛 Bug fix (non-breaking change that fixes an issue)

Changes Made

• hermes_cli/web_server.py: After proc.poll() returns a non-None exit code, call proc.wait(timeout=1) to reap the child process and .pop() the handle from _ACTION_PROCS. This prevents zombie accumulation for all dashboard actions (gateway-restart, hermes-update, etc.).
• tests/hermes_cli/test_web_server.py: Added test_finished_action_proc_is_reaped_and_removed — injects a mock FinishedProc into _ACTION_PROCS, calls the status endpoint, and asserts that proc.wait() was called and the entry was removed from the dict.

How to Test

1. Run the new test: pytest tests/hermes_cli/test_web_server.py::TestWebServerEndpoints::test_finished_action_proc_is_reaped_and_removed -xvs
2. Run related action tests: pytest tests/hermes_cli/test_web_server.py -k "action" -xvs
3. Verify no regressions: pytest tests/hermes_cli/test_web_server.py -x

Checklist

Code

• I've read the Contributing Guide
• My commit messages follow Conventional Commits (fix(scope):, feat(scope):, etc.)
• I searched for existing PRs to make sure this isn't a duplicate
• My PR contains only changes related to this fix/feature (no unrelated commits)
• I've run pytest tests/hermes_cli/test_web_server.py -q and all tests pass
• I've added tests for my changes (required for bug fixes, strongly encouraged for features)
• I've tested on my platform: macOS

Documentation & Housekeeping

• I've updated relevant documentation (README, docs/, docstrings) — or N/A
• I've updated cli-config.yaml.example if I added/changed config keys — or N/A
• I've updated CONTRIBUTING.md or AGENTS.md if I changed architecture or workflows — or N/A
• I've considered cross-platform impact (Windows, macOS) per the compatibility guide — or N/A
• I've updated tool descriptions/schemas if I changed tool behavior — or N/A

Code Intelligence

• Analyzed: hermes_cli/web_server.py:get_action_status() (caller: FastAPI route handler, flows: dashboard action status polling)
• Blast radius: LOW — localized to action status endpoint; no behavioral change for running processes
• Related patterns: _ACTION_RESULTS dict already handles completed non-spawned actions; this fix extends the same cleanup to spawned subprocesses

[alt-glitch]

Competing PR: #38040 also fixes #38032 and is broader in scope — it reaps the dashboard action zombies (same proc.wait() fix as here) plus the meet_bot.py pcm_pump leak and additional teardown in plugins/platforms/discord/adapter.py and tui_gateway/server.py. This PR is the minimal/focused variant (web_server.py only, with a dedicated reap test). Maintainers should pick one; see also #27073 / #35626 in the broader zombie-reap family.

[teknium1]

Superseded by #38049 (commit 78e2101), now merged to main — closes #38032. Your fix was the same approach (reap finished children via proc.wait() on poll); thanks for the clean, well-scoped patch. The merged version additionally migrates the exit code into _ACTION_RESULTS so repeat status polls stay consistent.