prospero-registration-helper
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided research titles and protocol text to generate PROSPERO registration forms using predefined templates.
- Ingestion points: User input for 'Title' and 'Protocol' is accepted as described in the workflow in
SKILL.mdand interpolated into templates inreferences/prompt_templates.md. - Boundary markers: The templates use
<>and{}delimiters to separate user-generated content from fixed system options, providing basic structural guidance. - Capability inventory: The skill executes a local Python script (
scripts/date_utils.py) to generate date strings; it does not possess network, file-write, or elevated system access. - Sanitization: No explicit sanitization or filtering of the user-provided text is performed before it is processed by the AI models.
- [COMMAND_EXECUTION]: The skill invokes a local Python script to calculate research timelines.
- Evidence:
SKILL.mdinstructs the agent to executepython scripts/date_utils.py. The script is a simple utility that uses the standarddatetimelibrary to return current and future dates, presenting no security risk.
Audit Metadata