skills/sapiom/skills/use-sapiom/Gen Agent Trust Hub

use-sapiom

Warn

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The tool 'sapiom_sandbox_exec' allows for the execution of arbitrary shell commands within persistent cloud sandboxes on the vendor's platform.
  • [REMOTE_CODE_EXECUTION]: The skill provides 'sapiom_run' and 'sapiom_job_deploy' tools to execute code snippets or scheduled tasks in multiple languages (Node.js, Python, Go, Rust) on remote infrastructure.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted data from the web.
  • Ingestion points: 'sapiom_fetch', 'sapiom_scrape', 'sapiom_crawl', 'sapiom_search', and 'sapiom_deep_search' fetch content from external URLs.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the documentation.
  • Capability inventory: The skill has high-privilege capabilities including remote code execution, shell access, and message publishing.
  • Sanitization: There is no documentation of sanitization or filtering of ingested content.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 3, 2026, 10:23 PM