The Agent Skills Directory

[REMOTE_CODE_EXECUTION] (CRITICAL): The skill exposes the /REST/nodes/{nodeName}/exec (POST) and /REST/nodes/nodeName/eval (GET) endpoints. These allow for the direct execution of arbitrary Python code and expressions on the host machine running the Nodel framework.
[COMMAND_EXECUTION] (HIGH): The skill provides tools for administrative persistence and destructive actions, specifically the /REST/nodes/{nodeName}/script/save endpoint for overwriting the node's logic and the /remove?confirm=true endpoint for deleting nodes.
[DATA_EXFILTRATION] (MEDIUM): The skill facilitates the retrieval of sensitive information, including full script source code via /REST/nodes/{nodeName}/script/raw and configuration parameters via /REST/nodes/{nodeName}/params, which may contain device credentials or internal network topology.
[PROMPT_INJECTION] (HIGH): The skill is highly vulnerable to Indirect Prompt Injection (Category 8):
Ingestion points: The agent is instructed to read console logs, activity logs, and files/contents (File: references/rest-api.md).
Boundary markers: Absent; there are no instructions to delimit or ignore instructions within these data sources.
Capability inventory: Extensive high-privilege capabilities including arbitrary Python execution (exec), script overwriting (script/save), and node deletion (remove) (File: SKILL.md).
Sanitization: Absent; the skill does not suggest any validation of the content retrieved from logs before processing or acting upon it.
Risk: An attacker-controlled external device could generate log messages containing instructions that the agent then executes using its RCE capabilities.

nodel-use