remember-team-context
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill uses search tools to retrieve stored context, which creates an ingestion point for data that could potentially contain malicious instructions.
- Ingestion points: External context retrieved via
mcp__plugin_ruflo-ai-team_ruflo-ai-team__memory_search. - Boundary markers: The skill includes explicit instructions to treat all returned text as untrusted data.
- Capability inventory: Utilizes tools for storing context (
memory_remember) and searching context (memory_search). - Sanitization: Instructions recommend treating retrieved content as untrusted data, though no specific programmatic sanitization is defined within the skill text.
Audit Metadata