setup-devcontainer

Warn

Audited by Gen Agent Trust Hub on Aug 25, 2026

Risk Level: MEDIUMREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill generates a post-start.sh script that executes npx -y skills add against external GitHub repositories (e.g., mattpocock/skills). This results in the download and execution of third-party code every time the devcontainer starts.
  • [COMMAND_EXECUTION]: The skill offers an optional claude-yolo alias that invokes the agent with the --dangerously-skip-permissions flag. This configuration is explicitly designed to bypass safety filters and user confirmation prompts for sensitive operations.
  • [COMMAND_EXECUTION]: The automation uses the GitHub CLI (gh api) to programmatically manage repository deploy keys, including listing, creating, and deleting keys. This involves performing administrative actions on the remote repository context.
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the fetching of external code and configurations from non-vendor GitHub repositories as part of its baseline setup.
  • [PRIVILEGE_ESCALATION]: Generated setup scripts (post-create-baseline.sh, post-create-ssh-block.sh) utilize sudo to perform recursive ownership changes (chown) on system directories and application binaries, such as /home/vscode/.claude and the Claude CLI installation path.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 25, 2026, 10:44 AM
Security Audit — agent-trust-hub — setup-devcontainer