[sighthound] Security findings in github/gh-aw

[github-actions]

Sighthound scan found 158 total findings, 120 actionable (outside test/testdata). Top 5 highest-severity findings (all Critical/High confidence, "Command Injection" pattern from exec.Command using pkgName/argument variables built from potentially dynamic data):

1. pkg/workflow/pip_validation.go:225 — exec.Command(uvPath, "pip", "show", pkgName, "--no-cache"). Remediation: ensure pkgName is validated against a strict allowlist pattern (e.g. PyPI package name regex) before use; never pass through shell interpolation; confirm args stay as a fixed slice (already true here) — the flag is likely due to pkgName originating from external config/frontmatter.

2. pkg/workflow/pip_validation.go:83 — exec.Command(pipPath, "index", "versions", pkgName, "--pre"). Remediation: validate pkgName against PyPI naming rules prior to invocation; same as above.

3. pkg/workflow/dependabot_manifests.go:263 — exec.Command(npmPath, "install", "--package-lock-only", "--ignore-scripts"). Remediation: verify working directory/manifest inputs are sanitized; confirm --ignore-scripts is always present (good practice) and that no untrusted path/args are concatenated.

4. pkg/cli/grype.go:336 — exec.Command(dockerPath, dockerArgs...). Remediation: audit construction of dockerArgs to ensure no unsanitized user-controlled strings (image names, tags) reach the slice; validate against expected formats.

5. pkg/cli/upgrade_command.go:507 — exec.Command(exe, newArgs...). Remediation: confirm exe and newArgs are derived only from trusted, internally-generated values (not user input) before re-exec.

General guidance: all flagged as "Command Injection" due to dynamic arguments passed to exec.Command. Since these use exec.Command (not sh -c), risk is present only if individual argument values are attacker-controlled without validation. Recommend auditing the upstream source of pkgName, dockerArgs, and newArgs in each case and adding explicit input validation/allowlisting where the values originate from external or less-trusted sources.

Full details: 158 findings scanned, 120 actionable; see workflow run for complete actionable.json.

Generated by 🛡️ Sighthound Security Scan · auto · 17.7 AIC · ⌖ 5.1 AIC · ⊞ 8.3K · ◷

[github-actions]

🍪 Issue Monster selected this for Copilot

I've identified this issue as a good candidate for automated resolution and requested assignment to the Copilot coding agent.

If assignment succeeds, the Copilot coding agent will analyze the issue and create a pull request with the fix.

Om nom nom! 🍪

🍪 Om nom nom by Issue Monster · gpt54 · 3.1 AIC · ⌖ 9.94 AIC · ⊞ 11.7K · ◷