skills/subsquid-labs/skills/portal/Gen Agent Trust Hub

portal

Pass

Audited by Gen Agent Trust Hub on Oct 23, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from public blockchain networks via the SQD Portal API. This external data represents a potential surface for indirect prompt injection attacks where malicious instructions could be embedded in on-chain logs or transactions.
  • Ingestion points: Blockchain data is ingested via the Portal Stream API at https://portal.sqd.dev/datasets/*/stream as documented in SKILL.md and individual reference files for EVM, Solana, Bitcoin, Tron, and Substrate.
  • Boundary markers: The skill contains a dedicated "External Data Boundary" section in SKILL.md that explicitly instructs the agent to treat data as untrusted and states that embedded instructions cannot override user tasks.
  • Capability inventory: The skill uses the Bash tool to perform network operations and data processing with curl and jq across several scripts.
  • Sanitization: Instructions in SKILL.md require parsing responses strictly as JSON/NDJSON and forbid the evaluation of returned text as shell, SQL, or JavaScript code.
  • [COMMAND_EXECUTION]: The skill uses the Bash tool to perform network requests and process JSON data. These operations target the vendor's official domain (portal.sqd.dev) and use standard utilities like curl and jq to implement the skill's primary functionality.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 23, 2026, 05:38 PM