amazon-sp-sales-traffic-analyzer
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data from Amazon Business Reports (JSON, CSV, or XLSX).
- Ingestion points: External data is ingested from API downloads (GET_SALES_AND_TRAFFIC_REPORT) or manual Seller Central exports as described in SKILL.md and processed by scripts/compute_metrics.py.
- Boundary markers: The skill explicitly defines a 'context-bloat boundary' in SKILL.md, instructing the agent to keep large source files outside the model context and only reason over compact metrics JSON artifacts.
- Capability inventory: The skill executes a local Python script (scripts/compute_metrics.py) and writes analysis reports (sales_traffic_analysis.md) to the local filesystem.
- Sanitization: The agent is instructed to introspect headers and perform data quality validation (e.g., currency matching, B2B subset validation) before processing data.
- [COMMAND_EXECUTION]: The skill uses a bundled Python script for deterministic data processing.
- Execution method: The agent executes scripts/compute_metrics.py via the command line with defined parameters such as input/output paths and analysis thresholds.
- Context: This command execution is localized to the provided script and is a core part of the skill's analytical function. The scripts do not perform network operations or attempt to access sensitive system files.
Audit Metadata