The Agent Skills Directory

[SAFE]: The skill contains informational documentation regarding the configuration of the agentmemory tool. It describes environment variables, port mappings, and feature toggles without executing commands or requesting elevated permissions.
[CREDENTIALS_SAFE]: The skill references sensitive configuration like ANTHROPIC_API_KEY and AGENTMEMORY_SECRET. It correctly suggests using a local .env file (~/.agentmemory/.env) for storage and uses placeholders for examples, which is a standard and safe practice for credential management.
[INDIRECT_PROMPT_INJECTION]: The skill documents the AGENTMEMORY_INJECT_CONTEXT feature, which enables automatic context injection into agent prompts. This represents an attack surface for indirect prompt injection at the platform level; however, the skill itself is purely configurational and does not implement the injection mechanism or provide malicious data payloads.
Ingestion points: The system reads memory data, potentially from external sources configured via AGENTMEMORY_URL or AGENTMEMORY_DATA_DIR (REFERENCE.md).
Boundary markers: Not specified in the configuration documentation.
Capability inventory: The configuration enables context injection and tool visibility (AGENTMEMORY_TOOLS).
Sanitization: Not specified in the configuration documentation.

agentmemory-config