right-to-rectification
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from data subjects, such as corrections and supporting evidence, which could contain malicious instructions designed to influence the agent's behavior during report generation or system updates.
- Ingestion points: Data enters the agent context through the "Receive and Log the Request" step in
SKILL.mdand through the--fieldand--subject-idarguments inscripts/process.py. - Boundary markers: The instructions in
SKILL.mdand the logic inscripts/process.pylack explicit delimiters or "ignore embedded instructions" warnings for the agent when handling user-supplied values. - Capability inventory: The skill includes capabilities to generate formal confirmation letters (
assets/template.md), update plans, and recipient notifications (scripts/process.py), all of which interpolate user-supplied data. - Sanitization: The
scripts/process.pyscript does not implement any validation, escaping, or sanitization of the strings provided via command-line arguments before they are incorporated into the JSON output and subsequent workflow logs.
Audit Metadata