The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill implements a persistent memory system that ingests external data including chat history, documentation, and codebases to build a Wiki and CodeGraph. This creates an attack surface where maliciously crafted files or code could influence the agent's future logic when the memory is recalled.
Ingestion points: SKILL.md describes features for indexing external code and documentation into a linked Wiki and CodeGraph.
Boundary markers: The skill does not provide specific instructions for using delimiters or 'ignore embedded instructions' warnings when processing this external content.
Capability inventory: The skill is granted access to the Bash and Write tools, which represents a capability tier that could be exploited if recalled memory contains malicious instructions.
Sanitization: There is no evidence of sanitization or validation performed on the ingested documents or code files before they are stored in the memory database.
[EXTERNAL_DOWNLOADS]: The skill configuration and setup process involve fetching source code and plugins from the vendor's infrastructure.
Evidence: Fetches the main repository from https://github.com/AnEntrypoint/agent-memory.git and installs the @tencentdb-agent-memory/memory-tencentdb plugin.

agent-memory