Security context
What an agent needs to avoid regressing past fixes and find the next vuln in this repo.
ProwlrBot/cyberbox
main @ 1fca61e
5
Fixes
0
CVEs
HIGH
Peak severity
50.0%
Coverage
Highlights
SSRF: 2 prior fixes. Scrutinize any change in this area.
caido-plugin/packages/backend/src/index.ts: most-fixed (2 issues). Treat as high-risk during review.
3 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.
SSRF: The backend component has repeatedly suffered from SSRF vulnerabilities where validators guarding endpoint resolution were bypassed, allowing unauthorized requests to internal resources. Developers must ensure all outgoing fetch requests and endpoint validations consistently apply strict host allowlists and correct URL parsing.
Command Injection: Direct interpolation of untrusted workflow inputs (such as 'bump') into shell step run commands allows arbitrary command execution within the runner context. Security must be maintained by always mapping untrusted inputs to intermediary environment variables instead of using raw string expansion.
Path Traversal: Path traversal during state operations allowed relative directory paths to resolve outside the intended state directory bounds. State-handling logic must strictly validate and canonicalize paths to reject relative steps like '..'.