rk-survey
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves ingesting and processing content from external research papers and abstracts, which constitutes an attack surface for indirect prompt injection.
- Ingestion points: The skill resolves papers from identifiers (DOIs, preprints), reads scientific notes, and processes full-text or abstract content from external sources and project files like
AGENTS.md(SKILL.md, references/citation-check.md). - Boundary markers: While the skill mandates a structured output format for the evidence map and search record, it does not provide explicit boundary markers or instructions to ignore potentially malicious commands embedded in the literature being reviewed.
- Capability inventory: The agent is permitted to read from the working directory and write structured research artifacts (checklists, records, and maps) to local files (SKILL.md, assets/).
- Sanitization: No specific protocols are defined for sanitizing, filtering, or escaping text extracted from external research papers before it is interpolated into summaries or evidence maps.
Audit Metadata