ai-citability-audit
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to fetch data, report summaries, and specific fix payloads from external endpoints hosted at blockquote.io and mcp.blockquote.io.
- [COMMAND_EXECUTION]: The skill provides explicit shell command templates using
curlto perform POST and GET requests to remote APIs, including the use of environment variables likeSCAN_ID. - [INDIRECT_PROMPT_INJECTION]: The skill establishes a workflow where untrusted data from an external API (recommendations and content/JSON-LD payloads) is ingested and acted upon by the agent.
- Ingestion points: API responses from
https://blockquote.io/api/v1/scan/andhttps://mcp.blockquote.io/mcpas described in SKILL.md. - Boundary markers: The skill provides no instructions for delimiting or verifying the external payloads before application.
- Capability inventory: The agent is instructed to "apply the smallest fix", which implies local file modification capabilities to update the audited page's metadata and content.
- Sanitization: No evidence of sanitization, validation, or escaping of the remote fix payloads is present in the skill instructions.
Audit Metadata