security-scanning
Pass
Audited by Gen Agent Trust Hub on May 9, 2026
Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill provides patterns for reading sensitive configuration files containing encryption keys.
- Evidence: A command in
SKILL.mdreads~/.config/sops/age/keys.txtto extract public keys for the sops encryption utility. - [PROMPT_INJECTION]: The skill processes untrusted local files and container images, creating an attack surface for indirect prompt injection.
- Ingestion points: Project directories, shell scripts, dependency lockfiles (e.g.,
package-lock.json), and container images. - Boundary markers: No delimiters or specific instructions are provided to separate untrusted data from the agent's processing instructions.
- Capability inventory: Shell command execution restricted to security utilities (
trivy,shellcheck,sops) via theBashtool. - Sanitization: Tool outputs are filtered and structured using
jqto minimize the injection surface.
Audit Metadata