skills/fatfingererr/macro-skills/analyze-high-unemployment-high-gdp-growth-fiscal-deficit-scenarios/Gen Agent Trust Hub
analyze-high-unemployment-high-gdp-growth-fiscal-deficit-scenarios
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS] (SAFE): The skill fetches economic data from FRED (https://fred.stlouisfed.org), a highly trustworthy government-affiliated source. The data is retrieved in CSV format using the requests library, and the implementation includes a local caching mechanism to minimize network requests.
- [COMMAND_EXECUTION] (SAFE): Workflows specify standard command-line operations for setting up a Python environment (pip install) and running local analysis scripts provided within the skill package. No usage of sudo or shell-injection-vulnerable commands was observed.
- [DATA_EXFILTRATION] (SAFE): Analysis is performed locally on the fetched economic data. The skill does not access sensitive user files (e.g., credentials, SSH keys) and has no functions to transmit data to untrusted third-party servers.
- [PROMPT_INJECTION] (SAFE): The workflows and reference files contain instructional language that is purely technical and process-oriented. There are no attempts to override agent safety protocols, reveal system prompts, or use jailbreaking techniques.
- [INDIRECT_PROMPT_INJECTION] (SAFE): While the skill ingests external data, the ingestion points are limited to structured numerical CSV data from FRED, which is processed via pandas. This methodology significantly reduces the risk of malicious instructions being successfully injected into the model's context through external data.
Audit Metadata