docs(#2634): document --agents flag in getting-started guide

[fullsend-ai-coder]

The getting-started guide for configuring GitHub did not mention the
--agents flag or explain the relationship between installed GitHub
Apps and active agents. Users who installed only a subset of apps expected that to automatically limit which agents were active.

Added a note after the GitHub Apps table clarifying that installing a subset of apps does not auto-limit agents — the --agents flag must be passed explicitly. Added a new "Enabling a subset of agents" subsection with a usage example (--agents triage,review) and a cross-reference to the full setup flags reference.

---

Closes #2634

Post-script verification

• Branch is not main/master (agent/2634-docs-agents-flag)
• Secret scan passed (gitleaks — f502487c5a393c3725515c8367ae924a8139ac84..HEAD)
• Pre-commit hooks passed (authoritative run on runner)
• Tests ran inside sandbox

[github-actions]

E2E tests did not run

E2E tests run automatically for org/repo members and collaborators on pull requests.

For other contributors, a maintainer must add the ok-to-test label after the latest push.

See E2E testing guide for details.

[github-actions]

E2E tests did not run

E2E tests run automatically for org/repo members and collaborators on pull requests.

For other contributors, a maintainer must add the ok-to-test label after the latest push.

See E2E testing guide for details.

[github-actions]

Site preview

Preview: https://c1d77103-site.fullsend-ai.workers.dev

Commit: 74d228373dcae6e7158098dc47c9a5beac23bd3c

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[fullsend-ai-retro]

🤖 Finished Retro · ✅ Success · Started 7:09 PM UTC · Completed 7:20 PM UTC
Commit: 74d2283 · View workflow run →

[fullsend-ai-retro]

Retro: PR #2660 — docs(#2634): document --agents flag in getting-started guide

Timeline

1. Issue #2634 filed by benoitf (2026-06-24 20:23 UTC) requesting docs improvements for the --agents flag and clarification that installing a subset of GitHub Apps does not auto-limit agents.
2. Triage triggered by ralphbean via /fs-triage (14:20 UTC). Agent completed by 14:25 UTC with a well-structured summary and recommended changes.
3. Code agent (run 28177300974) completed 14:26–14:36 UTC. Created PR #2660 with a clean 25-line docs-only change to docs/guides/getting-started/configuring-github.md.
4. No review agent feedback posted. A review run (28178425887) completed at 15:00 UTC but its artifact shows it reviewed PR feat(#2611): pin runners to ubuntu-24.04 with configurable __GH_RUNNER__ #2659 (runner pinning), not docs(#2634): document --agents flag in getting-started guide #2660. No initial review was dispatched for docs(#2634): document --agents flag in getting-started guide #2660. A second review was triggered only when ralphbean approved at 19:01 UTC, but the PR was merged 4 minutes later before it could finish.
5. Human approval by ralphbean at 19:01 UTC (empty body). PR merged at 19:05 UTC.

Assessment

Positive: The triage → code pipeline was fast (16 minutes end-to-end). The code agent produced a clean, well-scoped change that directly addressed the issue. The commit message and PR description were clear and well-structured. Single commit, single file, no rework needed.

Gaps identified:

• No review agent trace on this agent-authored PR — already tracked by #2637. Same pattern as PR fix(#2632): use SKILL.md frontmatter name for remote skill display #2633.
• Docs-only fast path for review — already tracked by #1900. A 25-line markdown addition is exactly the case that fast path is designed for.
• Duplicate e2e-gate comments — two identical 'E2E tests did not run' comments posted at the same timestamp (14:36:29Z). No existing issue covers this. Proposed below.

Proposals filed

• E2e-gate workflow posts duplicate comments on bot-authored PRs (in fullsend-ai/fullsend)