price-positioning
Pass
Audited by Gen Agent Trust Hub on Jul 14, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill architecture is self-contained and avoids all network operations, hardcoded credentials, or remote code dependencies.
- [COMMAND_EXECUTION]: The provided Python script (scripts/positioning.py) is implemented using only the standard library and performs no unsafe shell commands or dynamic code execution (eval/exec), ensuring safe local processing of analytical data.
- [PROMPT_INJECTION]: The skill manages the attack surface for indirect prompt injection through a mandatory evidence-labeling framework. 1. Ingestion points: The skill reads external CSV/JSON files via load_observations in scripts/positioning.py. 2. Boundary markers: Explicit instructions in references/evidence-and-honesty.md require labels such as [Data] and [Estimate] to scope external content. 3. Capability inventory: The skill has file-system read/write access but lacks network access or subprocess execution. 4. Sanitization: The script enforces strict normalization and type coercion (prices to floats) for all processed pricing data.
Audit Metadata