circuit-breaker-pattern
Pass
Audited by Gen Agent Trust Hub on Mar 18, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists of architectural instructions and logic definitions that do not contain malicious patterns, unauthorized data access, or hidden commands. The content is purely educational and functional for reliability engineering.- [NO_CODE]: The skill is implemented entirely in Markdown and YAML metadata. It does not include any executable scripts, binaries, or external library dependencies, significantly reducing the attack surface.- [COMMAND_EXECUTION]: The skill orchestrates the use of standard allowed tools (Bash, Grep, etc.) for resilience logic. It provides structured guidance on using alternative command-line tools when primary ones fail, which is a legitimate part of the described fault-tolerance pattern.- [INDIRECT_PROMPT_INJECTION]: The skill manages tool outputs, which constitutes a potential surface for indirect prompt injection. However, the logic includes mitigation steps such as explicit data labeling for quality and staleness (Step 6) and budget enforcement to prevent excessive error processing (Step 7). (Ingestion points: tool output strings; Boundary markers: data status labels in Step 6; Capability inventory: Read, Write, Edit, Bash, Grep, Glob; Sanitization: managed via state tracking and explicit manual fallback mechanisms).
Audit Metadata