skills/shyxin/skills/ssh-key-setup/Gen Agent Trust Hub

ssh-key-setup

Warn

Audited by Gen Agent Trust Hub on Jul 24, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes multiple shell commands to manage SSH infrastructure, such as ssh-keygen for key generation, chmod for securing file permissions, and ssh -T for verifying remote connectivity.
  • [CREDENTIALS_UNSAFE]: The skill accesses sensitive credential files including ~/.git-credentials and ~/.netrc. It uses cat and grep to search for legacy Personal Access Tokens (PATs) such as glpat. This pattern exposes cleartext authentication secrets to the AI agent during the execution phase.
  • [PROMPT_INJECTION]: The skill possesses surfaces for indirect prompt injection. 1. Ingestion points: The agent reads server metadata from references/server-facts.md and accepts arbitrary remote hostnames from user input. 2. Boundary markers: Absent; there are no explicit delimiters used to isolate untrusted data from instructions. 3. Capability inventory: The skill has the ability to read and write to the local filesystem and interact with the network via SSH. 4. Sanitization: None; data ingested from external sources or user input is passed directly to system commands without validation or escaping.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 24, 2026, 07:59 AM
Security Audit — agent-trust-hub — ssh-key-setup