para-memory-files
Pass
Audited by Gen Agent Trust Hub on Apr 18, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill presents an indirect prompt injection vulnerability by collecting user-provided content into memory files and later retrieving that content into the agent's context. \n
- Ingestion points: Daily notes in
$AGENT_HOME/memory/YYYY-MM-DD.mdand entity facts in$AGENT_HOME/life/entity folders. \n - Boundary markers: Absent; there are no instructions to delimit retrieved content or ignore embedded commands. \n
- Capability inventory: The agent uses the
qmdtool for retrieval and has write access to its own configuration files. \n - Sanitization: Absent; the skill does not specify any validation or filtering of stored information before it is re-integrated into the session context. \n- [COMMAND_EXECUTION]: The skill performs shell-based operations using the
qmdutility for semantic search, indexing, and retrieval of stored knowledge. \n- [COMMAND_EXECUTION]: The skill instructs the agent to persist lessons and behavior patterns by directly modifying its core configuration files, includingAGENTS.md,TOOLS.md, and other skill files. This self-modification capability allows for persistent changes to the agent's instructions based on data processed during sessions.
Audit Metadata