The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill configuration specifies downloading and executing the awslabs.agentcore-mcp-server package via uvx. This is the standard delivery method for the official AWS-labs toolset described in the skill and originates from a well-known service infrastructure.- [INDIRECT_PROMPT_INJECTION]: The skill features tools that ingest data from external documentation sources, which represents a potential attack surface. However, this is inherent to the skill's purpose and no malicious usage was found.
Ingestion points: search_agentcore_docs and fetch_agentcore_doc tools defined in SKILL.md.
Boundary markers: Absent.
Capability inventory: Includes system management tools like manage_agentcore_runtime and manage_agentcore_gateway.
Sanitization: Absent.- [SAFE]: No evidence of direct prompt injection, credential exfiltration, obfuscation, or persistence mechanisms was detected. The instructions follow standard AWS authentication and development practices.

aws-agentcore