monitor-inventor-mobility-signals-rd
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes patent data (titles, descriptions, applicant names) retrieved from external sources. This data is considered untrusted and could contain malicious instructions intended to manipulate the agent's behavior during the review process.\n
- Ingestion points: The
reviewed-data.jsonfile, which stores content retrieved from theadvanced_patent_searchandpatent_briefingtools.\n - Capability inventory: The
run_monitor.pyscript possesses file-writing capabilities to generate the HTML report.\n - Sanitization: The skill correctly utilizes
html.escapeto sanitize untrusted strings before they are rendered into the final HTML output.\n - Boundary markers: There are no explicit instructions or delimiters used when interpolating patent text into the rationale or technical relevance fields to prevent the agent from following embedded instructions.\n- [COMMAND_EXECUTION]: The documented workflow requires the agent to execute a local Python script,
scripts/generate_report.py, to transform the reviewed data into a human-readable report. This is an intended and documented function of the skill.
Audit Metadata