squid-perf
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local shell and Node.js scripts to perform its core functions, such as preflight.sh, fetch-logs.sh, parse.mjs, and report.mjs. It invokes the sqd CLI and expect for log automation. These executions are protected by restrictive regex validation of user-supplied deployment references to prevent shell injection.
- [INDIRECT_PROMPT_INJECTION]: The skill has a data ingestion surface as it processes log entries. 1. Ingestion points: fetch-logs.sh retrieves logs from Subsquid Cloud. 2. Boundary markers: None, but the parser uses regex to define line shapes. 3. Capability inventory: Shell execution of internal scripts and CLI tools; file system writes for reports. 4. Sanitization: Regex validation of deployment references, regex parsing in parse.mjs, and JSON encoding with script tag escaping in report.mjs. The ingestion is part of the primary vendor-related functionality.
- [SAFE]: Secret management is handled correctly by advising users to use the official sqd auth command with their own API keys rather than embedding credentials. Additionally, the skill employs environment variables to safely pass arguments into expect scripts, avoiding Tcl injection vulnerabilities.
Audit Metadata