Security context

What an agent needs to avoid regressing past fixes and find the next vuln in this repo.

msitarzewski/agency-agents

              main @ fc5a192
            

Fixes

CVEs

MEDIUM

Peak severity

Highlights

Command Injection: 1 prior fix. Scrutinize any change in this area.

.github/workflows/lint-agents.yml: most-fixed (1 issue). Treat as high-risk during review.

0 high-severity fixes in this history; regressions here are high-impact.

Recurring patterns

The bug types that recur here, drawn from past fixes, not open vulnerabilities.

Command Injection: Using direct interpolation of untrusted steps outputs in an inline shell script allows malicious file names or pull request payloads to execute arbitrary shell commands inside the runner context. Utilizing environment variables to pass these values prevents shell expansion bugs.