revnu-ai-readiness-audit
Pass
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill fetches and analyzes content from external, untrusted websites, which could potentially contain malicious instructions aimed at the AI agent.
- Ingestion points: Website content is retrieved via
curlinSKILL.mdandreferences/checks.mdbased on user-provided URLs. - Boundary markers: The skill explicitly instructs the agent to treat fetched text as data and to report rather than follow any instructions addressed to AI crawlers found on those pages.
- Capability inventory: The agent can execute
curl, run a Python-based HTML parser, and modify files within the repository if requested by the user. - Sanitization: Relies on instructional guardrails to ensure external content does not override agent behavior.
- [COMMAND_EXECUTION]: The skill uses shell commands, specifically
curl, and executes a local Python script provided in the reference files to extract metadata, headings, and JSON-LD structured data from audited pages. It may also attempt to execute a Python script from a sibling skill (revnu-ai-crawler-access) if installed. - [EXTERNAL_DOWNLOADS]: Fetches raw HTML and HTTP headers from external URLs provided by the user to perform its primary auditing function.
- [DATA_EXFILTRATION]: If requested by the user, the skill can send the target website's URL to
revnu.comto check for ongoing monitoring. This is a vendor-owned domain (Revnu) and is used for intended functionality.
Audit Metadata