predicate-logic
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
Bashtool to execute local Python scripts (scripts/z3_solve.pyandscripts/sympy_compute.py) for logic validation and formula simplification. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-supplied mathematical logic formulas and pass them as arguments to shell commands, creating a vulnerability surface if the agent does not properly escape the inputs.
- Ingestion points: User-provided predicate logic formulas are processed in the Decision Tree and Tool Commands sections of
SKILL.md. - Boundary markers: No specific delimiters or instructions (e.g., "ignore embedded instructions") are provided to the agent for handling user input within these commands.
- Capability inventory: The skill possesses
BashandReadcapabilities as defined in the YAML frontmatter and uses them to run complex scripts with string-interpolated arguments. - Sanitization: The instructions lack guidance on sanitizing or validating user-provided formulas before they are executed in a shell context.
Audit Metadata