admin-app-coolify

Pass

Audited by Gen Agent Trust Hub on Feb 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill performs extensive system administration via SSH on remote servers, which is consistent with its purpose as an infrastructure management tool.
  • Evidence:
  • Executes apt-get, systemctl, ufw, and docker commands to provision services.
  • Modifies /etc/ssh/sshd_config to configure secure root login settings.
  • Appends a generated public key to ~/.ssh/authorized_keys to allow the Coolify management container to communicate with the host via SSH.
  • [REMOTE_CODE_EXECUTION]: The skill utilizes remote script execution patterns to install core software components from official vendor repositories.
  • Evidence:
  • Downloads and executes the official Coolify installer: curl -fsSL https://cdn.coollabs.io/coolify/install.sh | bash.
  • Fetches the Cloudflare Tunnel agent (cloudflared) directly from the official Cloudflare GitHub releases.
  • Fetches the Docker GPG signing key from download.docker.com.
  • [EXTERNAL_DOWNLOADS]: Downloads necessary configuration files and binaries from trusted organizations and well-known services including Cloudflare, Docker, and CoolLabs.
  • [CREDENTIALS_UNSAFE]: A hardcoded string y^48ZTz3ZJ8J is present in references/TROUBLESHOOTING_CF1033.md.
  • Evidence:
  • The string is labeled as a "user selection" within a diagnostic report template. Given its length (12 characters) and context, it appears to be a dummy placeholder value rather than an active secret.
  • [PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection due to its ingestion of user-provided IPs, domains, and credentials.
  • Ingestion points: User parameters for server configuration defined in input-schema.json.
  • Boundary markers: SKILL.md mandates a manual confirmation checklist to verify parameters before any execution occurs.
  • Capability inventory: High-privilege capabilities including ssh remote command execution, docker orchestration, and firewall management.
  • Sanitization: input-schema.json provides basic validation using regex patterns for IP addresses and domain formats.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 25, 2026, 05:56 PM