gate-info-riskcheck
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted on-chain data which presents a risk of indirect injection.
- Ingestion points: Data enters the context through info_compliance_check_token_security and info_coin_get_coin_info tools as specified in SKILL.md.
- Boundary markers: The Report Template in SKILL.md interpolates external data fields (risk names, descriptions) directly into the prompt without structural delimiters.
- Capability inventory: The skill can execute local update scripts and call various read-only blockchain analysis tools.
- Sanitization: There is no documented sanitization of the retrieved blockchain metadata.
- [COMMAND_EXECUTION]: The skill uses local shell commands for environment detection and version management.
- SKILL.md contains a shell probe (Step 0) to determine routing between the legacy MCP workflow and a primary CLI tool.
- The skill includes and executes scripts/update-skill.sh and scripts/update-skill.ps1 for maintenance tasks.
- [EXTERNAL_DOWNLOADS]: Maintenance scripts fetch content from the official repository at github.com/gate/gate-skills. These downloads are used to update the skill files and are gated by a manual user confirmation step as required by the execution workflow in SKILL.md.
Audit Metadata