用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/dandye/adk_runbooks --skill hash-lookup-entity-chronicle命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
Use when evaluating threat detection opportunities (TDOs), generating synthetic UDM events, evaluating Chronicle rule coverage, and drafting YARA-L 2.0 rules.
Use when hunting for advanced persistent threat (APT) actor activity, tools, and infrastructure.
Use when hunting for lateral movement via PsExec, WMI, remote scheduled tasks, or WinRM.
正在显示 SKILL.md
| name | hash-lookup-entity-chronicle |
| description | Use when inspecting Chronicle entity records for known malicious or suspicious file hashes. |
| category | atomic |
| version | 1.0.0 |
| type | Skill |
| title | Skill: Atomic Runbook: Lookup File Hash Entity Activity in Chronicle |
| generated | {"by":"process:google-labs-jules","at":"2026-08-20T02:00:00Z"} |
ID: RB-ATOM-HASH-003
Version: 1.0
Last_Updated: 2025-05-30
Purpose: To retrieve a summary of a file hash's activity from Chronicle SIEM using the lookup_entity tool. This provides a quick overview of internal sightings (e.g., process executions, file observations), related alerts, and first/last seen times.
Parent_Runbook(s)/Protocol(s): rules-bank/indicator_handling_protocols.md#3-atomic-indicator-file-hash
Trigger: When a file hash requires an initial check for internal activity within Chronicle, often after external reputation checks.
file_hash: string - The file hash (MD5, SHA1, SHA256) to look up.
rb_hash_get_gti_report.md.hours_back (optional): integer - How many hours of historical data to consider. Defaults to 72 (3 days).secops-mcplookup_entityfile_hash (Input) to MCP Tool parameter entity_value.hours_back (Input or default) to MCP Tool parameter hours_back.lookup_entity.
rules-bank/mcp_tool_best_practices.md.chronicle_entity_summary: string - Textual summary from lookup_entity.related_alerts_count (optional, if parsable): integer.hosts_observed_count (optional, if parsable): integer - Number of unique hosts where hash was seen.output_status: string - ["Success", "NoInfoFound", "Failure"]output_message: string (if Failure).output_status is "Success":
chronicle_entity_summary.related_alerts_count > 0 OR hosts_observed_count > 0 (or summary indicates sightings) THEN
rb_hash_search_process_events_chronicle.md for detailed execution logs.output_status is "NoInfoFound":
output_status is "Failure"):
output_message.hours_back is 72. Adjust as needed.chronicle_entity_summary should look for "Observed on hosts:", "Related Alerts:", "File Executions:".output_status.rules-bank/ai_performance_logging_requirements.md)rules-bank/mcp_tool_best_practices.mdrules-bank/indicator_handling_protocols.mdrules-bank/ai_performance_logging_requirements.md