fix: scope full sync to GH_ORG and await Probot readiness

[decyjphr]

Why

Operators need to restrict full sync to a specific account instead of reconciling every account where the App is installed. This adapts #1053 to yadhav/fix-recent-issues, including its now-merged all-installations behavior from #1094.

Approach

• Export GH_ORG and select the matching repository-owning account case-insensitively. Fail clearly when no installation matches, including an empty installation list, rather than falling back to another account.
• When GH_ORG is unset, preserve the target branch's sequential sync of all repository-owning installations. Retain per-installation authentication/context, failure isolation, and aggregated { results, errors }.
• Log each selected account. Preserve enterprise enrichment and the existing info() slug lookup; enterprise installations are not repository sync targets.
• Return the scheduled sync promise and log/rethrow targeting or enumeration failures so node-cron reports failed execution. CLI errors retain nonzero exit status, including dry runs.
• Await Probot readiness before using its logger. Actual-runtime testing exposed a startup failure because Probot initializes asynchronously.
• Retain both targeting and fanout regression suites and smoke phases 21, 22, and 24. Update documentation to describe the combined behavior.

Validation

Node 22.12.0 / npm 10.9.0:

• After resolving the target-branch merge, npm run test:unit -- --runInBand --reporters=default: 889 passed, 12 skipped.
• Combined coverage includes unset/empty GH_ORG fanout, case-insensitive single-account targeting, enterprise exclusion, failure aggregation/isolation, targeted NOP configuration failure without fallback, CLI exit status, and cron failure reporting.
• Conflict-resolved CLI/test files pass targeted ESLint; syntax and whitespace checks pass.
• Six real-Probot CLI regressions, with synthetic credentials and networking blocked, originally failed before the readiness correction and passed afterward; they remain included in the combined suite.

Earlier live smoke, before this merge resolution: setup -> Phase 24 -> teardown passed 6 assertions with actual CLI completion and exit status verified. Independent cleanup confirmed unchanged original configuration, property definitions, admin policy, unrelated refs, and repository/team/role inventories. Owned processes exited and the port closed. No live smoke was rerun for the merge resolution.

Live coverage limitation: That smoke used a session-only verified-installation enumeration filter and hard authentication/API guards. It proves a contained single-test-org CLI/NOP run, not unfiltered live multi-account selection. Second-account selection and no-match behavior are covered offline. The first smoke attempt was rejected because Probot reloaded .env over explicitly blank execution flags; the retry verified effective flags after loading and restored the private environment copy exactly. No containment adapters or credentials are committed.

Earlier baseline validation found integration failures at the pre-existing Probot ESM import boundary (7 suites, 0 executed tests) and repository lint debt of 173 errors and 7 warnings. Those broad checks were not rerun for this merge resolution.

Provenance and integration

Originally based on 188c604929201774699fd1d8dabd445358cee9e8:

• Source 020145722fb1251f53d7331ccba6bdbc56217a60 from fix: sync the installation matching GH_ORG during full sync #1053 -> adaptation fd2d26e67213e280700fde3e4e983c7ad049a65f.
• Coupled CLI readiness correction -> a408d67184e44aed656303a71f11122f12e552e8.
• Merge resolution d8232af integrates target 4c5742b, including Fix repository and environment variable pagination #1093 and fix: sync all repository-owning app installations #1094.

The overlap with #1044's adaptation is now reconciled: all repository-owning installations when unset, only the matching account when GH_ORG is set. Combined local regression coverage passes; the earlier single-branch smoke is not presented as live validation of the merged implementation.

[Copilot]

Copilot review overview

🟢 Approval recommended

The implementation matches the stated targeting and fanout behavior with comprehensive regression coverage; only a minor documentation correction remains.

Review effort: Balanced
Findings: 1

Open (1)

• Replace incorrect “For e.g.” phrasing with “For example:” · New

What changed in this PR

Scopes full syncs using GH_ORG while preserving multi-installation fanout and ensuring Probot is ready before CLI use.

Changes:

• Adds case-insensitive installation targeting and clear failures.
• Preserves sequential fanout, enterprise enrichment, and cron error reporting.
• Expands CLI, unit, smoke, and documentation coverage.

File	Description
.env.example	Documents GH_ORG behavior.
README.md	Documents scoped full sync and smoke testing.
docs/​github-action.md	Updates GitHub Actions guidance.
full-sync.js	Awaits Probot readiness.
index.js	Implements targeting and cron rejection propagation.
lib/​env.js	Exports GH_ORG.
smoke-test.js	Adds targeted CLI smoke phase.
test/​unit/​full-sync.test.js	Tests real Probot CLI startup.
test/​unit/​lib/​env.test.js	Tests environment defaults and overrides.
test/​unit/​lib/​settings-batching.test.js	Updates Probot readiness mock.
test/​unit/​smoke-organization-targeting.test.js	Tests smoke-phase safeguards.
test/​unit/​sync-installation.test.js	Covers targeting, fanout, and failures.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.