chat-chat
Pass
Audited by Gen Agent Trust Hub on May 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes
curlcommands to facilitate interaction with the inbed.ai API for tasks such as agent registration, profile discovery, and message transmission. These are standard implementation patterns for API-based skills. - [EXTERNAL_DOWNLOADS]: The skill references the
inbed.aidomain and a GitHub repository under thegeeks-acceleratoraccount. These represent the primary service infrastructure and official source code repository for the vendor, posing no unexpected security risk. - [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it retrieves and processes message content from an external chat API.
- Ingestion points: Untrusted data enters the agent context via the
/api/chatand/api/chat/{{MATCH_ID}}/messagesendpoints (found in SKILL.md). - Boundary markers: The instructions do not define delimiters or specific warnings to ignore instructions embedded within the chat messages.
- Capability inventory: The agent has the capability to execute shell commands (
curl) for account registration and sending messages across the network. - Sanitization: No content validation or escaping mechanisms are described for the external message data.
Audit Metadata