akcli-schematic-review

Pass

Audited by Gen Agent Trust Hub on Sep 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from multiple external sources, creating a surface for indirect prompt injection attacks.
  • Ingestion points: The skill processes Altium (.SchDoc) and KiCad (.kicad_sch, .kicad_pcb) design files, Gerber folders, and external pinmap tables in CSV, JSON, and Devicetree (.dts) formats.
  • Boundary markers: The instructions do not define boundary markers or explicit delimiters to separate external hardware data from agent instructions, which could allow malicious metadata in the schematics to influence the agent.
  • Capability inventory: The skill utilizes akcli to perform complex analysis, structural checks, and power-tree verification, and writes findings to local JSON and Markdown files.
  • Sanitization: There is no evidence of sanitization or strict schema validation for input design data before the agent processes and summarizes it.
  • [COMMAND_EXECUTION]: The skill relies on the akcli CLI tool for all core operations, including structural ERC checks, electrical heuristics, and revision diffing. While appropriate for a hardware review skill, it requires the agent to execute multiple host-level commands to generate its report.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 7, 2026, 10:40 AM