The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill retrieves and injects content from the mem0_memory tool into the agent's context. While this data originates from past user interactions and could theoretically contain instructions, the skill functions as a read-only memory provider, which is its intended primary purpose.
Ingestion points: Data is ingested from mem0_memory search results (SKILL.md).
Boundary markers: No specific boundary markers or delimiters are defined for the output block to separate memory content from instructions.
Capability inventory: The skill is restricted to read-only search operations and context injection.
Sanitization: No explicit sanitization or filtering of retrieved memory content is performed before injection.
[SAFE]: The skill uses vendor-owned resources (mem0_memory tool) and follows standard practices for context management without exhibiting malicious patterns like obfuscation, persistence, or data exfiltration.

context-loader