ted-mosby
Warn
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFE
Full Analysis
- [EXTERNAL_DOWNLOADS] (MEDIUM): The skill instructs the user or agent to run
npm install -g ted-mosby. This installs a global package from an unverified source. Since the package is not from a trusted organization (e.g., Anthropic, Google, Vercel), its internal behavior cannot be verified through this manifest. - [CREDENTIALS_UNSAFE] (LOW): In the 'Check Prerequisites' section, the skill suggests running
echo $ANTHROPIC_API_KEY. Executing this command inside an agent's terminal session will print the full API key into the chat history and logs, leading to credential exposure. - [PROMPT_INJECTION] (LOW): Indirect Prompt Injection Surface. The skill processes untrusted external data (GitHub repositories via
ted-mosby generate -r https://github.com/user/repo). A malicious repository could contain comments or structured data designed to mislead the agent's architectural analysis. - Ingestion points: Repository files and remote GitHub URLs.
- Boundary markers: Not present in the command execution flow.
- Capability inventory: The tool performs file system reads and writes to generate documentation.
- Sanitization: Not specified in the skill description.
Audit Metadata