budget-pacing-monitor
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local Python utility script
ledger.pylocated within the plugin's directory structure at${CLAUDE_PLUGIN_ROOT}/scripts/connectors/ledger.py. This script is used to record and retrieve historical spend data for trend analysis. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external campaign data, which presents a surface for indirect prompt injection attacks.
- Ingestion points: The skill reads manually exported campaign report CSVs and GA4 traffic-acquisition exports provided by the user in the SKILL.md instructions.
- Boundary markers: The skill contains explicit instructions to treat all fetched or exported files as untrusted input and provides specific warnings to ignore any embedded instructions in data fields like campaign names or labels.
- Capability inventory: The agent has the ability to execute a local Python script (
ledger.py) and write pacing summaries to the localmemory/directory. - Sanitization: The instructions explicitly direct the agent to treat all exported values strictly as data and to avoid executing any content contained within the files.
Audit Metadata