skills/blackforge-so/skill/blackforge/Gen Agent Trust Hub

blackforge

Pass

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The scripts/check-catalog-sync.mjs script performs a network request to the vendor's API endpoint https://api.blackforge.so/v1/catalog to validate documentation accuracy against the live service.
  • [COMMAND_EXECUTION]: The skill uses scripts/latest-json.sh to wrap calls to the blackforge CLI, potentially installing it via npx using the official @blackforge-so/cli package.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests and interprets market data from the BlackForge API, which represents a potential attack surface if the source data were manipulated.
  • Ingestion points: Data retrieved via blackforge_catalog, blackforge_latest, and blackforge_series tool calls.
  • Boundary markers: Instructions explicitly mandate that the agent interprets all data strictly as measurements with units, though no formal prompt delimiters are specified for the tool outputs.
  • Capability inventory: The skill can execute local shell scripts and call MCP tools for network-based data retrieval.
  • Sanitization: Data is constrained to a predefined set of 120 metrics verified by the catalog tool.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 6, 2026, 07:08 AM
Security Audit — agent-trust-hub — blackforge