xss-stored
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONOBFUSCATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides a command example using
curlto allow users to verify if an injected canary string is persisted and rendered by a target application.\n- [OBFUSCATION]: The skill includes a URL-encoded payload example to illustrate how to bypass basic input filters that might block raw special characters.\n- [INDIRECT_PROMPT_INJECTION]: The skill documents the attack surface for stored script injection, where untrusted data is saved and later executed.\n - Ingestion points: Payloads are provided in
SKILL.mdfor injection into target application storage points such as profile fields and comments.\n - Boundary markers: The payloads are presented within markdown code blocks, but the skill does not include instructions to the agent to treat this data as untrusted or to ignore instructions embedded within the processed content.\n
- Capability inventory: The skill specifies the use of
curlfor interacting with web applications and inspecting their responses.\n - Sanitization: No logic is included for sanitizing the input payloads or the resulting output from the target servers.
Audit Metadata