understand-explain
Warn
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions direct the agent to use
grepwith user-supplied input:Grep to search the knowledge graph for the component: "$ARGUMENTS". This pattern is susceptible to command injection if the user provides shell metacharacters (e.g.,;,|,`) within the arguments, potentially allowing execution of arbitrary commands on the host system. - [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8) because it processes untrusted data to determine its next steps.
- Ingestion points: The agent reads data from
.understand-anything/knowledge-graph.json(SKILL.md). - Boundary markers: The instructions lack delimiters or explicit warnings to the agent to ignore any natural language instructions that might be embedded within the JSON summaries or tags.
- Capability inventory: The agent possesses the capability to execute shell commands (
grep) and read arbitrary files from the filesystem (SKILL.md). - Sanitization: No sanitization or validation logic is defined for the content extracted from the knowledge graph before it is used to perform file reads or generate explanations.
Audit Metadata