The Agent Skills Directory

[SAFE]: The skill acts as a proxy for Spamcop.net using the vendor's official domain (unbrowse.ai). All tool operations are restricted to the vendor's infrastructure and are intended for read-only data retrieval.
[EXTERNAL_DOWNLOADS]: The installation instructions utilize npx and refer to the vendor's own resources at unbrowse.ai and github.com/unbrowse-ai. These are legitimate vendor resources and do not represent a security risk.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external web pages (Spamcop.net), which is a common surface for indirect prompt injection.
Ingestion points: Data is fetched from external sources via the spamcop_net__get_w3m, spamcop_net__get_sc, and spamcop_net__read_page tools described in SKILL.md.
Boundary markers: The instructions in SKILL.md tell the agent to follow specific tool outcomes and only report results from successful runs, which acts as a behavioral constraint.
Capability inventory: The skill's capabilities are limited to making network requests to the vendor's specific MCP server. It does not have access to the local file system, arbitrary command execution, or persistence mechanisms.
Sanitization: The skill returns readable text and links from the target site, and the risk is mitigated by the lack of powerful capabilities available to the agent when using this skill.

spamcop