suanli-image-preheat
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses a local shell script (
scripts/call.sh) to executecurlcommands for API interaction. It also invokespython3to generate millisecond timestamps for request headers. These operations are restricted to the skill's primary function of API management. - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external API responses (cluster lists, task details, and resource availability) and integrates this information into the conversation. While this introduces an attack surface for malicious data from the API to influence the agent, the risk is inherent to the skill's purpose and is mitigated by the structured nature of the expected JSON responses.
- Ingestion points: Data is received via
scripts/call.shfromopenapi.suanli.cnendpoints. - Boundary markers: None explicitly defined in the scripts; the agent is instructed to extract specific fields like
task_idandstatus. - Capability inventory: Execution of shell scripts, network requests via
curl, and reading local payload files. - Sanitization: None observed in the provided shell scripts; the skill relies on the agent's internal safety guardrails for processing API data.
Audit Metadata