roam-research-pro
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data retrieved from external Roam Research graphs through tools like
get_blockanddatalog_query, creating an attack surface where malicious instructions embedded in the graph data could influence agent behavior during write operations. - Ingestion points: Data returned by MCP tools
get_blockanddatalog_queryas referenced inSKILL.mdandreferences/code-fence.md. - Boundary markers: The skill lacks explicit instructions for the agent to use delimiters or ignore embedded instructions within the ingested graph data.
- Capability inventory: The skill utilizes
create_page,create_block,update_block, anddatalog_queryMCP tools for modifying the graph environment. - Sanitization: No sanitization, escaping, or validation logic is provided to handle potentially malicious content within the retrieved block strings before they are processed or re-written.
Audit Metadata