abaqus-optimization

Pass

Audited by Gen Agent Trust Hub on Oct 29, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONOBFUSCATION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest user-defined optimization goals (stiffness, volume, stress) and translate them into configuration changes using privileged tools.
  • Ingestion points: User-provided values for objectives, constraints, and manufacturing parameters defined in SKILL.md.
  • Boundary markers: None identified; the instructions do not specify the use of delimiters or 'ignore' instructions for interpolated data.
  • Capability inventory: The skill has access to Write, Edit, and a scoped Bash(abaqus:*) tool to modify the environment and execute commands.
  • Sanitization: No evidence of input validation or escaping logic to prevent malicious input from influencing the resulting configuration or shell execution.
  • [COMMAND_EXECUTION]: The skill uses the Bash(abaqus:*) tool to interact with the Abaqus environment. While the tool is scoped, the instructions allow for dynamic configuration based on user input, which could lead to command injection if input is not properly handled by the agent.
  • [OBFUSCATION]: The file UPSTREAM.md contains a Byte Order Mark (BOM, U+FEFF) at the start of the file. This is a common artifact from certain text editors but is technically a zero-width character that can be used to bypass simple text-based security filters.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 29, 2026, 02:14 AM