model-debugging

SUSPICIOUS: the skill’s core purpose is legitimate and its install sources are mostly benign, but its actual footprint is broader than a normal debugging guide. It reads raw secret files, forwards bearer tokens to APIs, exposes an embedded token in the documentation, accesses user-level telemetry, and includes production-changing actions like SCP and service restarts. This is not confirmed malware because data flows go to official services and first-party hosts, but the credential exposure and operational scope make it a medium-high risk skill.

This module is a straightforward monitoring/triage script that queries Tinybird for 403 error patterns. No clear signs of malware (persistence, reverse shells, local data harvesting, or payload execution) are present. The primary security issues are (1) sensitive credential handling—an admin token is transmitted in the curl URL, which can leak via logs/process listings/telemetry—and (2) potential SQL/query injection into the remote Tinybird SQL layer via unescaped TIER_FILTER (and broadly unvalidated HOURS/MIN_ERRORS). Treat as low-to-moderate security risk and mitigate by validating inputs and using safer authentication/query practices (avoid token in URL; parameterize/escape or restrict tier values).