space-xhs-title

Pass

Audited by Gen Agent Trust Hub on Jul 31, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill consists entirely of natural language instructions and markdown formatting. No executable code, scripts, or system-level commands are present.
  • [PROMPT_INJECTION]: The skill includes explicit safety boundaries and 'Red Lines' that instruct the AI to avoid generating deceptive or non-compliant content, reinforcing platform safety standards rather than bypassing them.
  • [DATA_EXFILTRATION]: There are no indicators of network activity, external tool calls, or attempts to access sensitive local data such as credentials or configuration files.
  • [REMOTE_CODE_EXECUTION]: The skill does not contain instructions for downloading, installing, or executing external software or remote scripts.
  • [INDIRECT_PROMPT_INJECTION]: While the skill processes user-supplied text (drafts and themes), it lacks any exploitable capabilities like file system access or network operations that could be triggered by malicious input.
  • Ingestion points: Processes user-provided themes and drafts in SKILL.md.
  • Boundary markers: Relies on natural language context rather than specific technical delimiters.
  • Capability inventory: No dangerous tools, subprocess calls, or external communication capabilities identified.
  • Sanitization: Input is not sanitized at the technical level, but the output is filtered by the skill's own compliance logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 31, 2026, 06:44 PM
Security Audit — agent-trust-hub — space-xhs-title