threat-triage
Pass
Audited by Gen Agent Trust Hub on Mar 23, 2026
Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill identifies and scans for sensitive infrastructure components, including SIEM platforms (Splunk, Microsoft Sentinel), EDR/XDR configurations, and STIX/TAXII server settings.
- [DATA_EXFILTRATION]: The skill accesses the sensitive directory path
~/.claude/projects/to append execution telemetry to a local log file. - [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection due to the ingestion of untrusted data. Ingestion points: processes user-supplied arguments and local configuration files discovered during analysis. Boundary markers: absent; the skill does not use specific delimiters to separate ingested data from agent instructions. Capability inventory: writes analysis reports to the filesystem and appends metadata to telemetry files in sensitive directories. Sanitization: absent; no validation or escaping of analyzed data is specified.
Audit Metadata