The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon data from external project documents and search tool results, which constitutes an attack surface for instructions embedded within that data.
Ingestion points: The agent is instructed to read project documentation (docs/*.sdoc) and retrieve search results from Magic Context and claude-mem tools.
Boundary markers: There are no instructions to use delimiters or specific safety warnings to prevent the agent from following malicious instructions found within the retrieved text.
Capability inventory: The skill possesses extensive capabilities to search, expand context, and write/save memory (ctx_search, ctx_expand, ctx_memory, ctx_note, claude_mem_search, migrate-mem, save-mem), which could be misdirected by poisoned memory content.
Sanitization: The instructions do not define any validation or filtering mechanisms for the external content before it is integrated into the agent's context.

load-mem