metaprev
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute the
@forsvn/metaprevutility usingnpx. This involves running shell commands that interact with the local file system (writing HTML reports) and network. - [EXTERNAL_DOWNLOADS]: The skill relies on fetching the
@forsvn/metaprevpackage from the NPM registry at runtime. While the package is associated with the skill's author (forsvn-labs), it represents an external dependency execution pattern. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process metadata (OpenGraph/Twitter tags) from arbitrary external URLs. This creates a surface where a malicious website could embed instructions in its meta tags to influence the agent's behavior when it parses and reports on the 'validated' content.
- Ingestion points: The tool fetches HTML content from user-provided or environment-detected URLs (e.g.,
npx @forsvn/metaprev <url>). - Boundary markers: The output is structured into terminal tables, JSON, or a local HTML file. The instructions advise the agent to "review every value" and treat snippets as a "starting point."
- Capability inventory: The skill can execute shell commands, write files to the local disk (using the
-oflag), and perform network fetches. - Sanitization: The tool parses tags via regex and provides a "repair brief," but the agent's interpretation of the resulting text remains a vulnerability surface if the source site is adversarial.
Audit Metadata