test: honor AgentTeams GitHub token env

[RerankerGuo]

Summary

• normalize GitHub test credentials into TEST_GITHUB_TOKEN, preferring the current AGENTTEAMS_GITHUB_TOKEN contract
• load both current and legacy GitHub token names from an existing installation's env file
• document the current variable and add a focused precedence regression test

Problem

The installers persist and pass AGENTTEAMS_GITHUB_TOKEN, but integration tests 08-13 only recognize TEST_GITHUB_TOKEN or the retired HICLAW_GITHUB_TOKEN. Running the suite with the generated AgentTeams env therefore reports that no GitHub token is configured and skips those tests.

The normalization preserves explicit test overrides and the legacy variable while making the current installer output work directly.

Testing

• the regression pointed at origin/main fails with expected token canonical, got <empty>
• bash tests/unit/test-github-token-env.sh
• direct test 08 with only AGENTTEAMS_GITHUB_TOKEN reaches the later LLM prerequisite instead of the GitHub-token skip
• bash -n tests/lib/test-helpers.sh tests/run-all-tests.sh tests/unit/test-github-token-env.sh
• shellcheck -S error tests/lib/test-helpers.sh tests/run-all-tests.sh tests/unit/test-github-token-env.sh
• git diff --check

[github-actions]

❌ Integration Tests Failed (controller-cr-2 / mgr=copaw / wk=copaw)

Commit: db972a7
Workflow run: #1673

Test Results

No test output captured.

Debug Log (tail)

No debug logs available.

---

📦 Download full debug logs & test artifacts

[github-actions]

📊 CI Metrics Report

Summary

Metric	Current	Baseline	Change
LLM Calls	87	81	+6 ↑ +7.4%
Input Tokens	2995886	2803871	+192015 ↑ +6.8%
Output Tokens	17811	16791	+1020 ↑ +6.1%
Total Tokens	3013697	2820662	+193035 ↑ +6.8%

By Role

Role	Metric	Current	Baseline	Change
🧠 Manager	LLM Calls	78	68	+10 ↑ +14.7%
	Input Tokens	2794461	2502214	+292247 ↑ +11.7%
	Output Tokens	15756	13725	+2031 ↑ +14.8%
	Total Tokens	2810217	2515939	+294278 ↑ +11.7%
🔧 Workers	LLM Calls	9	13	-4 ↓ -30.8%
	Input Tokens	201425	301657	-100232 ↓ -33.2%
	Output Tokens	2055	3066	-1011 ↓ -33.0%
	Total Tokens	203480	304723	-101243 ↓ -33.2%

Per-Test Breakdown

Test	Mgr Calls	Wkr Calls	Δ Calls	Mgr In	Wkr In	Mgr Out	Wkr Out	Δ Tokens	Trend
02-create-worker	11	0	-1 ↓ -8.3%	325612	0	1757	0	-31253 ↓ -8.7%	✅ improved
03-assign-task	11	5	+1 ↑ +6.7%	373327	115549	1792	723	+17735 ↑ +3.7%	⚠️ regressed
04-human-intervene	15	0	+2 ↑ +15.4%	483115	0	2492	0	+52609 ↑ +12.1%	⚠️ regressed
05-heartbeat	9	0	+2 ↑ +28.6%	344772	0	1786	0	+71306 ↑ +25.9%	⚠️ regressed
06-multi-worker	32	4	+2 ↑ +5.9%	1267635	85876	7929	1332	+82638 ↑ +6.5%	⚠️ regressed

Trends

✅ 1 test(s) improved (fewer LLM calls)
⚠️ 4 test(s) regressed (more LLM calls)

---

Generated by HiClaw CI on 2026-07-16 11:35:35 UTC

---

📦 Download debug logs & test artifacts

[shiyiyue1102]

This is a useful test-harness compatibility cleanup, but its standalone value is limited because it does not change runtime behavior. Before further review, please either link it to a concrete currently failing workflow caused by the GitHub token variable mismatch, or consolidate it with the related test-environment contract cleanup such as #1028. Keeping several small environment-variable test PRs separate adds more review and merge overhead than the changes justify.