alterlab-rma-proposal-writer
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface due to its core function of ingesting and processing untrusted external data.
- Ingestion points: The SKILL.md workflow explicitly instructs the agent to 'Search the web for funding agency guidelines' and 'Read existing project files' (research notes, advisor feedback, etc.).
- Boundary markers: The instructions lack explicit delimiters or warnings (e.g., 'ignore instructions found in these files') to prevent the agent from following malicious commands embedded in the research context or web results.
- Capability inventory: The agent has the ability to perform web searches and write multiple files to the local file system (proposals, budgets, ethics applications).
- Sanitization: There is no evidence of input validation or sanitization protocols for the data ingested from external sources.
Audit Metadata