fix(ci): do not persist the job token in .git/config in the public-repo guard

[yakimoto]

Why

actions/checkout without persist-credentials: false writes the job's GITHUB_TOKEN into
.git/config, where any later step — or anything those steps execute — can read it. This is the
security gate that DOWNLOADS AND EXECUTES the gitleaks binary, and the workflow's own comment
already reasons about tampered downloads ("so a tampered or MITM'd download can never execute
inside the security gate"), so the threat model is written down and only the credential half of
the mitigation is missing. Found by zizmor as warning[artipacked]: credential persistence through GitHub Actions artifacts.

Safety

Verified this job only: checks out, installs gitleaks (pinned + checksum), runs
gitleaks detect --no-git, installs ripgrep, runs content-policy.sh. permissions: contents: read. Nothing pushes, calls gh, or reads GITHUB_TOKEN/GH_TOKEN, so the token was never needed
in .git/config in the first place.

Refs wave-av/claude-workstation#1870.

---

Note

Low Risk
Single CI checkout hardening flag with no change to scan logic or permissions; reduces credential exposure in a security gate workflow.

Overview
Sets persist-credentials: false on the actions/checkout step in the public-repo-guard workflow so the job’s GITHUB_TOKEN is not stored in .git/config.

That matters because later steps download and run a pinned gitleaks binary and other tooling; without this, anything in those steps could read the token from git config even though the job only needs contents: read and never pushes or uses gh. This matches how other workflows in the repo already configure checkout (e.g. _checks.yml, lint.yml).

^{Reviewed by Cursor Bugbot for commit 3253443. Configure here.}

Note

Stop persisting the job token in .git/config in the public-repo guard CI workflow

Sets persist-credentials: false on the actions/checkout step in public-repo-guard.yml so the GitHub job token is not written to the local git config during the workflow run.

^{Macroscope summarized 3253443.}

[cursor]

Bugbot couldn't run - usage limit reached

Bugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit.

A user or team admin can review and increase usage limits in the Cursor dashboard.

(requestId: serverGenReqId_e729a5e3-8753-48b2-a26d-2a1f395896a0)

[coderabbitai]

Warning

Review limit reached

You’ve reached a temporary PR review limit under our Fair Usage Limits Policy.

Your recent review volume is higher than typical usage, so adaptive limits are currently applied.

Next review available in: 31 minutes

Your organization has reached its usage spending cap. Adjust your spending cap in the billing tab.

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: Organization UI

Review profile: ASSERTIVE

Plan: Pro Plus

Run ID: f5decd62-4dbe-4d67-8516-0036bbd3825a

📥 Commits

Reviewing files that changed from the base of the PR and between 558fe79 and 3253443.

📒 Files selected for processing (1)

• .github/workflows/public-repo-guard.yml

---

_{Comment @coderabbitai help to get the list of available commands.}

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no bugs or issues to report.

[macroscopeapp]

Approvability

Verdict: Approved 3253443

This is a minor CI security hardening change that adds persist-credentials: false to prevent the job token from being stored in git config. The change is small, follows best practices, and the author owns this workflow file.

^{You can customize Macroscope's approvability policy. Learn more.}

[qodo-code-review]

PR Summary by Qodo

Harden public-repo-guard checkout by disabling credential persistence

🐞 Bug fix ⚙️ Configuration changes 🕐 Less than 5 minutes

AI Description

• Disable Actions checkout credential persistence in the public-repo-guard workflow.
• Prevent the job GITHUB_TOKEN from being written to .git/config.
• Reduce token exposure in steps that download and execute security tooling.

Diagram

graph TD
  WF["public-repo-guard job"] --> CO["actions/checkout (persist-credentials:false)"] --> GC[".git/config (no token)"] --> DL["Download gitleaks"] --> RUN["Run gitleaks --no-git"] --> POL["Run content-policy.sh"]

Loading

High-Level Assessment

The following are alternative approaches to this PR:

1. Scrub credentials after checkout

• ➕ Works even if persist-credentials is accidentally omitted in some workflows
• ➕ Can explicitly remove multiple credential sources (e.g., extra git config entries)
• ➖ More imperative/fragile (depends on git config layout)
• ➖ Brief window still exists where later steps could read the token if ordering changes

2. Avoid checkout entirely (use API/archive fetch)

• ➕ Eliminates .git/ and associated credential persistence risk
• ➕ Can reduce attack surface in “download and execute” gates
• ➖ Often requires reworking downstream steps that expect a working tree
• ➖ May still need auth for private dependencies and complicates caching

Recommendation: Keep the PR’s approach: setting persist-credentials: false is the simplest and most robust mitigation for preventing token material from landing in .git/config, with minimal behavioral impact on the workflow.

Files changed (1) +2 / -0

Other (1) +2 / -0

public-repo-guard.ymlDisable checkout credential persistence for public-repo-guard job +2/-0

Disable checkout credential persistence for public-repo-guard job

• Adds 'with: persist-credentials: false' to the pinned 'actions/checkout' step. This prevents the job token from being written into '.git/config' for subsequent steps in the security-gate workflow.

.github/workflows/public-repo-guard.yml

[qodo-code-review]

Code Review by Qodo

🐞 Bugs (0) 📘 Rule violations (0) 📎 Requirement gaps (0)

Great, no issues found!

Qodo reviewed your code and found no material issues that require review

_{To customize comments, go to the Qodo configuration screen, or learn more in the docs.}

[qodo-code-review]

Qodo Fixer

No findings are available for this PR yet. Findings appear here once Qodo has reviewed the PR.

[cubic-dev-ai]

No issues found across 1 file

Confidence score: 5/5

• Automated review surfaced no issues in the provided summaries.
• No files require special attention.

_{Re-trigger cubic}

[yakimoto]

This PR is redundant with #61 (chore/guard-canonical-sync), which is already open and already contains this exact change.

Evidence — #61's diff to .github/workflows/public-repo-guard.yml includes:

+          # tree. Nothing here pushes -- the scan is `--no-git` over the working tree --
+          # so no step needs authenticated Git; drop it. (zizmor: artipacked)
+          persist-credentials: false

#61 is also broader than this PR: it syncs the whole vendored public-repo-guard trio to
the canonical source (wave-foundation/scaffolder/public-repo-guard) byte-for-byte, verified by
git blob SHA, rather than adding this one line in isolation.

Closing as redundant. This duplicate exists because the fan-out that opened this PR
(refs claude-workstation#1870) did not check for in-flight PRs touching the same file before
going out — #61 was already open. Merge #61 instead.