story
Warn
Audited by Gen Agent Trust Hub on Jun 26, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill performs network requests to GitHub (api.github.com) to check for the latest version of the toolkit from the repository worldwonderer/oh-story-claudecode.
- [REMOTE_CODE_EXECUTION]: The skill includes an update procedure that prompts the user to execute 'npx skills add worldwonderer/oh-story-claudecode -y -g'. This command downloads and installs code from a repository that is not associated with a trusted organization or the provided author context, potentially leading to untrusted code execution.
- [COMMAND_EXECUTION]: The skill utilizes several CLI tools including curl, gh, npx, grep, and jq for version monitoring and update management.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it reads untrusted data from the local environment to determine its state. Ingestion points: The skill scans local project directories (checking for specific subdirectories) and reads the content of a .active-book file. Boundary markers: Absent. Capability inventory: Shell command execution (curl, npx, gh) and skill routing. Sanitization: No validation or sanitization is performed on directory names or file contents before they are processed by the agent.
Audit Metadata