silent-cli
Pass
Audited by Gen Agent Trust Hub on Feb 21, 2026
Risk Level: SAFE
Full Analysis
- Prompt Injection (SAFE): No instructions attempting to override agent behavior or bypass safety filters were detected. The content is purely instructional regarding CLI tool usage.
- Data Exposure & Exfiltration (SAFE): No hardcoded credentials or sensitive data were found. Examples utilize standard placeholders such as 'xxx', 'password', or 'ghp_xxx'. The skill documents environment variables used for authentication (e.g., PGPASSWORD, VAULT_TOKEN) as part of its intended purpose for CI/CD automation.
- Obfuscation (SAFE): The markdown files contain clear, readable text. No base64, zero-width characters, or homoglyphs are present.
- Unverifiable Dependencies & Remote Code Execution (SAFE): The skill does not include or download executable code. It documents standard package management commands (e.g., apt-get install, npm install) and provides examples for data transfer tools like curl and wget, which are expected given the skill's scope.
- Privilege Escalation (SAFE): Commands involving sudo are documented for system management tools (e.g., ufw, nft, tcpdump). These are appropriate for the specific administrative tools being described.
- Indirect Prompt Injection (LOW): The skill documents tools that ingest untrusted data (e.g., curl, wget) and have high-privilege capabilities (e.g., ssh, terraform, docker).
- Ingestion points: curl.md, wget.md, git.md.
- Boundary markers: Absent in provided command examples.
- Capability inventory: ssh.md (remote execution), terraform.md (infrastructure management), docker.md (container orchestration).
- Sanitization: Not applicable as the skill is a reference guide. The author mitigates risk by including explicit 'WARNING' blocks for security-sensitive flags like 'StrictHostKeyChecking=accept-new' or '--noconfirm'.
Audit Metadata