Aloudata CAN SKILLS - anomaly-detection

The name/description and the SKILL.md consistently describe a metric anomaly-detection role. The skill delegates all data fetching to a separate metric-query skill (so it does not request API keys itself) and focuses on baseline building, statistical tests, and reporting — which is appropriate for the stated purpose.

The instructions stay within anomaly-detection scope (clarify scope, choose baselines, compute statistics, report results). They explicitly delegate all API/auth work to metric-query. The SKILL.md includes executable Python snippets that use numpy; standard-model agents are told to copy/execute these blocks. This is functionally normal for an analysis skill but means the agent will execute code locally (and requires numpy in the runtime) — not a security mismatch but an operational consideration.

No install spec and no code files — instruction-only. This is the lowest-risk install profile (nothing is downloaded or written to disk by the skill itself).

The skill declares no required environment variables or credentials. It relies on metric-query to perform authenticated data access; therefore any credentials used will come from that other skill, which is coherent for this skill's scope. Verify the trustworthiness of metric-query since it handles auth.

always=false and there is no indication the skill requests permanent agent-wide privileges or modifies other skills' configs. Autonomous invocation is permitted by platform default but does not combine here with other red flags.