[t3dotgg]

Stacked PRs were evicting the shared CI cache. A PR whose base is not main saves a ~1 GB pnpm store that only its own reruns can read.

Measured from Blacksmith logs and cache inventory:

• 145 full Linux x64 pnpm stores were saved between two Mac releases. 126 of them came from stacked PRs, 16 from default-base PRs, and 3 from main pushes.
• 19 stacked PRs held 17.71 GiB of the 24.55 GiB inventory. Those entries used only 2 distinct keys, so most were duplicate copies.
• Example: this Linux job on a stacked PR restored a fallback store, downloaded 1 package, and then uploaded a 1,001,159,631-byte copy.

Fix: Save pnpm store now runs only when the event is not pull_request, or the PR base is the repo's default branch. Stacked PRs still restore and install as before. Pushes, scheduled, dispatched and called releases, and default-base PRs save as before. The filtered/full-store rule and the lockfile verification cache do not change.

Verification: I evaluated the real if: expressions with @actions/expressions for 15 cases (stacked and default-base PRs, push, schedule, dispatch, hits, misses, filtered installs, and tooling-only installs), 75/75 checks passed. Against origin/main the only differences are the 3 stacked-PR store saves. vp run check:changed passes.

The savings are modeled, not measured yet: if the 126 stacked-PR saves had not happened, ~126 GB less would have been uploaded in that window, and the ~17.7 GiB would have stayed free for shared entries.

Council review

Rounds: 1

Delegate	Round 1
Claude Opus 5.5	9/10
Muse Wonder Full	9/10
Codex GPT-6.1 Sol	9.5/10

🤖 Generated with Claude Code using Claude Opus 5.5.

[macroscopeapp]

Approvability

Verdict: Approved at c486f0b

Macroscope's review found this PR approvable — Stacked PRs stop uploading pnpm stores while retaining cache restoration and dependency installation. The change is confined to disposable CI caching; default-base PRs and non-PR workflows retain their existing behavior.

^{You can add or adjust custom eligibility rules. Learn more.}

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	5.0 KiB	5.0 KiB	0 B (0.0%)	6.8 KiB	✅
Codex	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Codex	Live turn WebSocket wire	1.2 KiB	1.2 KiB	0 B (0.0%)	2.0 KiB	✅
Codex	Live turn WebSocket decoded	21.5 KiB	21.5 KiB	0 B (0.0%)	24.4 KiB	✅
Codex	Live turn messages	1	1	0 (0.0%)	8	✅
Claude	Total thread wire	5.0 KiB	5.0 KiB	0 B (0.0%)	6.8 KiB	✅
Claude	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Claude	Live turn WebSocket wire	1.2 KiB	1.2 KiB	0 B (0.0%)	2.0 KiB	✅
Claude	Live turn WebSocket decoded	21.9 KiB	21.9 KiB	0 B (0.0%)	24.4 KiB	✅
Claude	Live turn messages	1	1	0 (0.0%)	8	✅

Baseline: 86d8838 · PR result: c486f0b · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 108.5 KiB
• Claude decoded thread snapshot: 108.8 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

[coderabbitai]

No actionable comments were generated in the recent review.

Walkthrough

The pnpm store save step now excludes pull requests targeting non-default branches. Other events and pull requests targeting the default branch remain eligible under the existing conditions.

Changes: pnpm Store Cache

Layer / File(s)	Summary
Update store save eligibility .github/actions/setup-ci-deps/action.yml	The save-step condition excludes pull requests targeting non-default branches. The updated comment describes the behavior and the existing conditions.

Estimated code review effort: 1 (Trivial) | ~5 minutes

Change: Bug fix

Suggested reviewers: yash-singh1

Merge Risk: Minimal · up to c486f

Stacked pull requests no longer save pnpm stores; no merge-blocking risk was identified.

Pre-merge checks: 4

Check name	Status	Explanation
Title check		The title clearly describes the change to pnpm cache saves for stacked PRs.
Linked Issues check		Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check		Check skipped because no linked issues were found for this pull request.
Description check		The description covers the problem, change, and verification. It is mostly complete, but lacks the required Scope and approval section and an approval link or exemption rationale.

Finishing Touches

Generate unit tests (beta)

• Commit to this branch
• Create a new PR

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Recent review info

Run configuration

• Source mode: Auto
• Repository: pingdotgg/t3code
• Branch: t3/skip-stacked-pr-cache-saves
• File: .coderabbit.config.ts
• Commit: c486f0b
• Layers: Repository TS config
• Review profile: CHILL
• Plan: Team
• Run ID: 12b62833-e32c-4afd-a973-b6d653aa7fd4

Commits

Reviewing files that changed from the base of the PR and between 86d8838 and c486f0b.

Files selected for processing: 1

• .github/actions/setup-ci-deps/action.yml

Included review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 5 remain after this review.