skills/dcoon/agent-kit/wiki-memory/Gen Agent Trust Hub

wiki-memory

Pass

Audited by Gen Agent Trust Hub on Oct 27, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill is a productivity tool for maintaining state in a markdown file (wiki/memory/MEMORY.md) using a local Obsidian MCP server. It does not perform network requests, access sensitive system credentials, or include any obfuscated code.
  • [INDIRECT_PROMPT_INJECTION]: The skill reads and acts upon rules and lessons stored in its memory file, which constitutes an ingestion point for potentially untrusted or instructions.
  • Ingestion points: Reads instructions from wiki/memory/MEMORY.md.
  • Boundary markers: The skill does not define explicit delimiters to separate data from instructions in the memory file.
  • Capability inventory: Capabilities are limited to vault operations via an MCP server; shell commands are explicitly user-gated.
  • Sanitization: No input validation or content sanitization is described before the agent processes recorded rules.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 27, 2026, 01:49 PM