modern-web-guidance
Warn
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill's core functionality relies on downloading the 'modern-web-guidance' package from the NPM registry at runtime using 'npx'. The package is not provided by a well-known service or a trusted organization, and it does not match the naming patterns associated with the skill's author ('anis-dr').
- [COMMAND_EXECUTION]: The instructions command the agent to run 'npx -y modern-web-guidance@latest search' and 'npx -y modern-web-guidance@latest retrieve'. This executes arbitrary code from a remote package on the user's local system to perform task-critical functions.
- [INDIRECT_PROMPT_INJECTION]: The skill establishes an attack surface where the agent processes external data without safety boundaries.
- Ingestion points: Data retrieved from the 'modern-web-guidance' CLI tool (e.g., search results and markdown guides).
- Boundary markers: Absent. The agent is not instructed to treat the retrieved content as untrusted or to ignore embedded instructions.
- Capability inventory: The agent has shell execution capabilities (npx) and file manipulation capabilities.
- Sanitization: Absent. The agent directly incorporates retrieved text into its reasoning and code generation process.
- [PROMPT_INJECTION]: The skill uses behavioral override markers such as 'MANDATORY: Execute FIRST' and 'Do NOT skip' to force the agent to prioritize the skill's instructions over its default safety or operational guidelines for web development tasks.
Audit Metadata