Security context

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

griptape-ai/griptape

              main @ ebf422b
            

Fixes

CVEs

CRITICAL

Peak severity

50.0%

Coverage

Highlights

Privilege Escalation: 4 prior fixes. Scrutinize any change in this area.

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

1 high-severity fix 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: Command execution via subprocess.Popen allowed potential shell injection if unsanitized inputs were passed to the shell. Hardening required disabling shell execution globally and parsing parameters via shlex.split.

Privilege Escalation: GitHub Actions workflows historically lacked explicit least-privilege definitions, exposing default write tokens that could be abused if an attacker compromised a dependency or runner.

Auth Bypass: LLM agents could invoke unauthorized internal tool methods in memory components unless restricted by a default activity allowlist.