mid-engagement-ir-detection
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides Bash monitoring scripts that use
curlto probe target endpoints andgrepto analyze local operation logs for lockout patterns. - [DYNAMIC_EXECUTION]: A utility script uses a shell heredoc to execute inline Python code for comparing current metrics against established baselines.
- [INDIRECT_PROMPT_INJECTION]: The monitoring scripts ingest and process data from external HTTP responses and account lockout logs.
- Ingestion points: HTTP response timings, headers, and sizes captured from external targets; login status codes from
o365_results.jsonl. - Boundary markers: Absent; the scripts rely on specific field extraction from structured or semi-structured data.
- Capability inventory: Local file read/write, shell command execution (
curl,grep), and inline Python execution. - Sanitization: Numerical data is sanitized via
bc,cut, and Python'sint()conversion, which mitigates injection risk from the target responses.
Audit Metadata