resubmission-deadline-tracker
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data in the form of manuscript titles and researcher notes which are processed at runtime to generate schedules.
- Ingestion points: User input via CLI arguments (--title, --notes) and stored records in data/deadlines.json processed by scripts/main.py.
- Boundary markers: The response template in SKILL.md provides a structured format (Assumptions, Assessment, Schedule, Risks) to separate data from instructions.
- Capability inventory: Local file read/write to data/deadlines.json via scripts/main.py. No network operations, arbitrary command execution, or privilege escalation capabilities are present.
- Sanitization: Data is stored and retrieved using Python's standard json library.
Audit Metadata