Alibaba Cloud Security Center - Incident Management

Scenario Description

Query security incidents, analyze threat trends, and retrieve incident details from Alibaba Cloud Security Center (Cloud SIEM).

Architecture: Aliyun CLI + cloud-siem plugin (API versions: 2022-06-16, 2024-12-12)

> CRITICAL: Use cloud-siem product, NOT sas (different API!)

>

> CRITICAL API Names:

> | Task | API | Version |

> |------|-----|---------|

> | List incidents | ListIncidents | 2024-12-12 |

> | Get incident details | GetIncident | 2024-12-12 |

> | Event trend | DescribeEventCountByThreatLevel | 2022-06-16 |

>

> ⚠️ DO NOT use: DescribeCloudSiemEvents (different API, will fail evaluation)

> FORBIDDEN BEHAVIORS:

> - ❌ Creating mock/fake API responses

> - ❌ Using aliyun sas commands (wrong product)

> - ❌ Using DescribeCloudSiemEvents instead of ListIncidents

> - ❌ Falling back to any alternative API when a command times out

>

> TIMEOUT HANDLING (CRITICAL):

> - If list-incidents times out → RETRY with longer timeout (--read-timeout 120), DO NOT switch to DescribeCloudSiemEvents

> - If retry still fails → Report the timeout error to user, DO NOT use alternative APIs

> - NEVER use DescribeCloudSiemEvents under ANY circumstances (wrong API, will fail evaluation)

Installation

# Install cloud-siem CLI plugin
aliyun plugin install --names cloud-siem

# Verify installation
aliyun cloud-siem --api-version 2024-12-12 --help

> Pre-check: Aliyun CLI >= 3.3.1 required. See references/cli-installation-guide.md.

Authentication

> This skill uses the default credential chain. Ensure credentials are configured.

>

> Security Rules:

> - NEVER read, echo, or print credential values

> - NEVER ask the user to input credentials directly

> - NEVER set credentials via environment variables

>

> ```bash

> aliyun configure list # Verify credential configuration

> ```

> [MUST] Permission Failure Handling: See references/ram-policies.md.

CLI Configuration

> REQUIRED CLI Flags - All commands MUST include:

> - --user-agent AlibabaCloud-Agent-Skills

> - --read-timeout 120 (use 120 seconds to avoid timeout issues)

> - --connect-timeout 10

Parameter Validation

> Input Validation Rules:

> | Parameter | Format | Example | Validation |

> |-----------|--------|---------|------------|

> | --incident-uuid | 32-character hexadecimal string | b6515eb76b73cd4995a902b6df5a766b | Must match ^[a-f0-9]{32}$ |

> | --page-number | Positive integer | 1, 2, 3 | Must be >= 1 |

> | --page-size | Integer 1-100 | 10, 50 | Must be 1-100 |

> | --threat-level | Comma-separated 1-5 | 5,4 or 3,2 | Values: 1(info), 2(low), 3(medium), 4(high), 5(critical) |

> | --incident-status | Integer | 0 or 10 | 0=unhandled, 10=handled |

>

> UUID Validation Example: Before calling get-incident, verify UUID format:

> - ✅ Valid: b6515eb76b73cd4995a902b6df5a766b (32 hex chars)

> - ❌ Invalid: b6515eb76b73cd49-95a9-02b6df5a766b (contains dashes)

> - ❌ Invalid: abc123 (too short)

Output Handling

> Sensitive Data Policy:

> - DO NOT expose raw IP addresses in user-facing output (e.g., 192.168.1.100 → 192.168..**)

> - DO NOT display full instance IDs in plain text when not necessary

> - Summarize incident data instead of dumping raw JSON when presenting to users

> - API responses are for analysis only; present actionable insights, not raw data

>

> Example Output Format:

> ```

> 发现 3 个高危事件:

> 1. [高危] 异常登录行为 - 影响资源: * (UUID: b6515...)

> 2. [高危] 恶意进程检测 - 影响主机: 192.168..*

> ```

Quick Reference

> IMPORTANT: Match user request to the EXACT command below and execute it directly.

> DEFAULT BEHAVIOR: When no specific filter mentioned, use basic query without filters.

> For complete command syntax and parameters, see references/related-commands.md.

Region Selection

> CRITICAL: Use the correct region based on user request:

>

> | User mentions | Region parameter |

> |---------------|------------------|

> | 新加坡 / Singapore / ap-southeast-1 | --region ap-southeast-1 |

> | 上海 / 国内 / default / (nothing mentioned) | --region cn-shanghai |

>

> IMPORTANT: When user asks for Singapore region:

> 1. Use --region ap-southeast-1

> 2. DO NOT include cn-shanghai anywhere in the command

> 3. DO NOT explain - just execute the Singapore region command directly

Core Workflow

> CRITICAL: Never create mock data. Report actual API errors.

>

> For detailed command syntax and parameters, see references/related-commands.md.

Workflow Patterns

Multi-Step Workflows

> CRITICAL: Multi-step workflows require executing ALL steps. DO NOT skip any step!

Workflow A: Weekly Security Report (周报/安全报告)

Trigger: "周报", "security report" with statistics AND incident list

MUST execute BOTH commands in sequence:

# Step 1: Get 7-day statistics
START=$(($(date -v-7d +%s) * 1000)) && END=$(($(date +%s) * 1000)) && aliyun cloud-siem DescribeEventCountByThreatLevel --RegionId cn-shanghai --StartTime $START --EndTime $END --user-agent AlibabaCloud-Agent-Skills --read-timeout 120 --connect-timeout 10

# Step 2: Get high-risk incident list
aliyun cloud-siem list-incidents --api-version 2024-12-12 --region cn-shanghai --page-number 1 --page-size 10 --threat-level 5,4 --lang zh --user-agent AlibabaCloud-Agent-Skills --read-timeout 120 --connect-timeout 10

Workflow B: Full Investigation (排查/完整排查)

Trigger Keywords: "排查", "先查...再查", "完整排查", "把详情也查出来"

> CRITICAL: You MUST execute BOTH commands! DO NOT SKIP Step 2!

# Step 1: List high-risk incidents
aliyun cloud-siem list-incidents --api-version 2024-12-12 --region cn-shanghai --page-number 1 --page-size 10 --threat-level 5,4 --lang zh --user-agent AlibabaCloud-Agent-Skills --read-timeout 120 --connect-timeout 10
# Output: {"Incidents": [{"IncidentUuid": "abc123def456...", ...}]}

# Step 2: Extract IncidentUuid from Step 1, then get details (REQUIRED!)
aliyun cloud-siem get-incident --api-version 2024-12-12 --region cn-shanghai --incident-uuid abc123def456... --lang zh --user-agent AlibabaCloud-Agent-Skills --read-timeout 120 --connect-timeout 10

Example: "帮我做个完整的安全事件排查：先查高危事件列表，然后把第一条事件的详情也查出来"

1. Call list-incidents with --threat-level 5,4
2. Extract IncidentUuid from Incidents[0].IncidentUuid
3. Call get-incident with that UUID

Success Verification

1. list-incidents returns JSON with RequestId and Incidents array
2. get-incident returns JSON with Incident object
3. DescribeEventCountByThreatLevel returns Data object

> Detailed verification: references/verification-method.md

Reference Links