AI-generated analysisPublished automatically and not human-verified. Validated context appears in community notes below.
← Watch feed
Low 47 Bitcoin

ci: gen PR labels from master instead of fork

Public commit record

What the developer wrote

Authored by satsfy (Renato Britto)

85/100 · Strong
ci: gen PR labels from master instead of fork

The labeler checked out the fork and ran cargo metadata on
its manifests while holding the base repo token, a github
actions security vulnerability called pwn request.

But we don't need the fork checkout to generate labels, this
commit now does it from master.
✓ Descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides detailed explanatory context✓ Names security-relevant behavior explicitly
The short version

What changed, and why it matters

This commit fixes a GitHub Actions workflow that was checking out code from a pull request's fork while holding a write-capable repository token. That setup is a known 'pwn request' attack pattern: a malicious fork could run code during the workflow and potentially steal or misuse the repository token. The fix removes the fork checkout and generates labels using only the trusted master branch, so untrusted code no longer runs with privileged credentials.

Recommended action

Verify that no other workflows in the repository check out refs/pull/<number>/merge or fork refs while using the default GITHUB_TOKEN, and consider pinning all actions to commit SHAs as already done here. Review the gen_label_config.sh script to ensure it does not itself invoke network or file operations that could reintroduce risk.

Security signals we found

01

Workflow executed untrusted PR merge code with base-repo GITHUB_TOKEN

02

Removal of refs/pull/<number>/merge checkout

03

Self-described fix for 'github actions security vulnerability called pwn request'

04

Reduction of token exposure surface by no longer checking out fork code

Risk score

Why this scored 47/100

Our methodology →
Potential impact 12/30
Exploitability 10/25
Stealth signal 8/15
Affected reach 6/15
Confidence 7/10
Evidence quality 4/5
Human-validated context

Community notes

Notes can correct, qualify, or add evidence to the AI analysis. Every note shown here has been validated by a human moderator.

No validated notes yet.

The AI analysis stands alone for now. Submit a note if you can add evidence or important context.