submit-to-agentlaunch
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the Bash tool to perform API interactions using
curl. These operations include retrieving directory metadata, submitting agent JSON payloads, and performing upvotes. - [EXTERNAL_DOWNLOADS]: The skill retrieves JSON data from the Agent Launch API hosted on
agents-launch.lovable.app. This includes fetching directory rules, enums, and lists of existing agents. - [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface.
- Ingestion points: Data is ingested from the public Agent Launch directory through metadata queries and agent list retrievals (SKILL.md).
- Boundary markers: The instructions lack delimiters or specific directives to ignore potential command-like text within the ingested directory data.
- Capability inventory: The skill has access to network tools (Bash/WebFetch) and can perform state-changing operations via POST requests (SKILL.md).
- Sanitization: There is no evidence of sanitization or structural validation to ensure that ingested descriptions do not influence the agent's subsequent logic or tool usage.
Audit Metadata