dev-to
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill configuration and MCP server connections target unbrowse.ai, which is the official domain for the skill's author. These connections are necessary for the skill's functionality to replay dev.to API requests server-side.
- [INDIRECT_PROMPT_INJECTION]: The skill acts as a bridge to search results and content from dev.to. This ingestion of external, user-generated data creates an attack surface where malicious instructions could be embedded in forum posts or comments.
- Ingestion points: Data enters the agent context through the search results returned by the
dev_to__post_anonymous_2tool and theunbrowse.scrapeorunbrowse.maptools (SKILL.md). - Boundary markers: The skill instructions do not specify the use of delimiters or warnings for the agent to ignore instructions within the retrieved content.
- Capability inventory: The skill provides tools for browser actions (
unbrowse.browse.act), task execution (unbrowse.run), and generic data fetching on the target domain (SKILL.md). - Sanitization: No explicit sanitization or filtering of the dev.to content is described in the provided skill files.
Audit Metadata