acrobatics
Pass
Audited by Gen Agent Trust Hub on Apr 27, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill is entirely metaphorical and instructional. It provides a logical methodology for the agent to follow when managing brittle sequencing or API integrations without invoking any dangerous commands or scripts.
- [NO_CODE]: No code, scripts, or binaries are included in the skill. Its functionality is restricted to prompt-based guidance for the AI agent's reasoning process.
- [PROMPT_INJECTION]: The skill processes user-supplied workflows, representing an attack surface for indirect prompt injection.
- Ingestion points: User-defined workflows and integration sequences described during the /acrobatics invocation (SKILL.md).
- Boundary markers: Absent, although the procedure instructs the agent to define 'no-touch boundaries.'
- Capability inventory: None; the skill does not include or request any specific tools or file access.
- Sanitization: No input sanitization or escaping mechanisms are specified.
Audit Metadata