The Agent Skills Directory

[METADATA_POISONING]: The skill includes deceptive metadata regarding its origin. The SKILL.json file identifies the author as 'Anthropic Marketplace 3p' and includes tags such as 'official-marketplace' and 'official', which contradicts the actual author identity ('420company'). This impersonation of an official entity may lead users to misjudge the skill's provenance.
[EXTERNAL_DOWNLOADS]: The skill communicates with the official Bright Data API (api.brightdata.com) to retrieve session logs. This interaction is consistent with the skill's stated purpose and targets a well-known service.
[INDIRECT_PROMPT_INJECTION]: The skill analyzes external data from API responses, creating a surface for potential indirect prompt injection attacks.
Ingestion points: Untrusted data, such as target URLs and error messages from the Bright Data API, is ingested into the agent context through the logic described in SKILL.md.
Boundary markers: No explicit delimiters or boundary markers are defined to isolate the external API data from the agent's instructions.
Capability inventory: The skill's capabilities are limited to making network GET requests; it does not have access to higher-risk operations like local file system writes or arbitrary command execution.
Sanitization: There is no evidence of sanitization or validation logic applied to the external API content before it is processed by the agent.

brd-browser-debug