signals-scout-anomaly-detection
Pass
Audited by Gen Agent Trust Hub on Jun 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a bundled Python script (
scripts/ks2.py) and uses Python heredocs to perform Kolmogorov-Smirnov statistical tests. The execution involves fetching the file from the skill bundle, writing it to a temporary directory, and invoking the Python interpreter. The script is localized to the skill package and does not perform network or sensitive file system operations. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from external analytics queries and interpolates it into user-facing descriptions and notebooks.
- Ingestion points: Data is retrieved via the
insight-query,execute-sql, anddashboard-insights-runtools, which can return attacker-influenced content such as insight names, dashboard titles, or event property values. - Boundary markers: The instructions do not define clear delimiters or include warnings to ignore instructions embedded in the analyzed data.
- Capability inventory: The skill has the capability to write to the platform via
notebooks-createandsignals-scout-emit-signal, potentially misleading users if the content is manipulated. - Sanitization: No explicit sanitization or validation of the retrieved analytics data is implemented before its interpolation into prompt templates.
Audit Metadata