blackforge
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions and scripts (such as
scripts/latest-json.sh) facilitate the execution of theblackforgeCLI tools to retrieve market snapshots and series data. - [EXTERNAL_DOWNLOADS]: The skill dynamically fetches and executes official vendor packages
@blackforge-so/mcpand@blackforge-so/cliusingnpx. It also performs network requests toapi.blackforge.soto synchronize the metrics catalog and fetch market data. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted market data originating from various external cryptocurrency exchanges via the BlackForge API.
- Ingestion points: External data enters the agent context through the
blackforge_catalog,blackforge_latest,blackforge_series, andblackforge_symbolstools described inSKILL.md. - Boundary markers: The instructions do not define explicit boundary markers or delimiters to wrap external data, though they provide strong behavioral guidelines for the agent to treat outputs as measurements.
- Capability inventory: The skill possesses the capability to execute shell commands and perform network operations via the provided CLI wrappers and MCP tools.
- Sanitization: There is no evidence of automated sanitization or filtering applied to the external data fields (like labels or metadata) before presentation to the agent.
Audit Metadata