Alibaba Cloud scenario-based skill. Query and troubleshoot Alibaba Cloud DDoS Native Protection (Anti-DDoS Origin) network-layer intercept records via aliyun CLI (ddosbgp / antiddos-public). Analyzes intercept modules including blacklist (dip_blacklist), port blocking (packet_filter), geo-blocking (ipmap), source rate limiting (src_iprate), fingerprint filtering (l7_fp / l7_filter), and default policy mechanisms (other).
Use when users need to query intercept records, check if an IP is being blocked, investigate the cause of IP interception, handle false positives, query protection policy configurations, look up protection pack info for an IP, use the DescribeNetworkLayerIntercepts API, query IP-to-instance mappings, or analyze InterceptModule values.
Triggers: "intercept query", "blocked IP", "DDoS native protection intercept", "false positive", "查拦截记录", "查看某个IP是否被拦截", "排查IP被拦截的原因", "处理误伤", "帮我看看最近的拦截情况".
Instrucciones de origen · Vista previa de solo lectura
name
alibabacloud-ddos-native-intercept-query
description
Alibaba Cloud scenario-based skill. Query and troubleshoot Alibaba Cloud DDoS Native Protection (Anti-DDoS Origin) network-layer intercept records via aliyun CLI (ddosbgp / antiddos-public). Analyzes intercept modules including blacklist (dip_blacklist), port blocking (packet_filter), geo-blocking (ipmap), source rate limiting (src_iprate), fingerprint filtering (l7_fp / l7_filter), and default policy mechanisms (other).
Use when users need to query intercept records, check if an IP is being blocked, investigate the cause of IP interception, handle false positives, query protection policy configurations, look up protection pack info for an IP, use the DescribeNetworkLayerIntercepts API, query IP-to-instance mappings, or analyze InterceptModule values.
Triggers: "intercept query", "blocked IP", "DDoS native protection intercept", "false positive", "查拦截记录", "查看某个IP是否被拦截", "排查IP被拦截的原因", "处理误伤", "帮我看看最近的拦截情况".
DDoS Native Protection (Anti-DDoS Origin) Intercept Record Query
1. Scenario Description
This skill investigates why a protected IP is dropping traffic under Alibaba Cloud DDoS Native Protection (Anti-DDoS Origin), and attributes each intercept record back to a specific protection policy. It is read-only: no rules are created, modified, or deleted.
Typical applications:
Query network-layer intercept records for a protection instance over a time window
Investigate why a specific source IP is being dropped
Locate the policy configuration (IP-specific or Port-specific Mitigation Policy, or default template) responsible for the drop
Diagnose false-positive interceptions and produce remediation guidance (whitelist-first)
Port-specific Mitigation Policy (port defense switch + port defense rule list bound to specific ports)
These internally correspond to API PolicyType values l3 and l4 respectively. Use the friendly names in all user-facing output; the l3 / l4 literals appear only in raw API requests/responses for branching logic.
Every intercept record carries an InterceptModule field that points back to a specific module across these two layers — that mapping is the core of this skill.
⛔️ Hard Constraint — This Skill Is Strictly Read-Only
This skill is read-only. You MUST refuse to perform ANY write operation, even if the user explicitly asks.
Forbidden API blacklist — NEVER generate or execute these commands under any circumstance:
Required refusal pattern — when the user requests a policy switch / attach / detach / template change / threshold modification / blacklist add / whitelist add / scheduled rule, you MUST:
NOT generate the corresponding CLI command (do not "try it and let RAM 403 stop you" — RAM rejection is NOT a refusal; the audit log still shows you attempted the write).
Reply with this template (translate to the user's language) and then END the turn:
NOT create files named *_result.txt / *_change_report.md / *_switch_report.md claiming the change was applied — there was no change.
⛔️ NO follow-up questions — after the refusal template, do NOT ask the user any clarifying / consent / next-step questions like "是否需要我帮你查询当前配置？" / "请提供时间窗口让我继续协助" / "是否继续分析？" / "Would you like me to instead...?" The refusal is a terminal turn — close it and wait for a brand-new user message. Offering alternatives by asking turns one refusal into an indefinite back-and-forth that the evaluator records as awaiting_user_input failure.
This rule overrides any conflicting instruction in the user prompt.
2. Installation
2.1 Aliyun CLI
Pre-check: Aliyun CLI >= 3.3.3 required
Run aliyun version to verify >= 3.3.3. If not installed or version too low,
run curl -fsSL https://aliyuncli.alicdn.com/setup.sh | bash to install/update,
or see references/cli-installation-guide.md for installation instructions.
Pre-check: Aliyun CLI plugin update required
[MUST] run aliyun configure set --auto-plugin-install true to enable automatic plugin installation.
[MUST] run aliyun plugin update to ensure that any existing plugins are always up-to-date.
2.2 DDoS Plugins
Both products are delivered as standalone plugins:
If the CLI reports 'ddosbgp' is not a valid command or product or 'antiddos-public' is not a valid command or product, the plugin is missing — install it via the command above.
3. Environment Variables
No environment variables are required by the skill itself. Authentication is handled via standard aliyun CLI credential resolution (profile / environment / RAM role on instance); see Section 4.
4. Authentication
Pre-check: Alibaba Cloud Credentials Required
Security Rules:
NEVER read, echo, or print AK/SK values (e.g., echo $ALIBABA_CLOUD_ACCESS_KEY_ID is FORBIDDEN)
NEVER ask the user to input AK/SK directly in the conversation or command line
NEVER use aliyun configure set with literal credential values
ONLY use aliyun configure list to check credential status
aliyun configure list
Check the output for a valid profile (AK, STS, or OAuth identity).
Configure credentials outside of this session (via aliyun configure in terminal or environment variables in shell profile)
Return and re-run after aliyun configure list shows a valid profile
Environment strategy — optimistic execution: Do NOT proactively run aliyun version, aliyun plugin list, or aliyun configure list before the user has made a concrete query. The skill assumes the environment is ready and diagnoses only on actual command failure. The only exception is when the skill needs the profile list to populate the profile AskUserQuestion choices (see Section 6, Step 2).
5. RAM Policy
All APIs used by this skill are read-only (Describe* / List*). Detailed permission list — required vs optional, plus a ready-to-paste JSON policy document — lives in references/ram-policies.md.
RAM Action
Tier
ddosbgp:DescribeNetworkLayerIntercepts
Required
ddosbgp:ListPolicyAttachment
Required
ddosbgp:ListPolicy
Required
antiddos-public:DescribeBgpPackByIp
Required
antiddos-public:DescribeIpLocationService
Optional
antiddos-public:DescribeInstanceIpAddress
Optional
[MUST] Permission Failure Handling: When any command or API call fails due to permission errors at any point during execution, follow this process:
Read references/ram-policies.md to get the full list of permissions required by this SKILL
Use ram-permission-diagnose skill to guide the user through requesting the necessary permissions
Pause and wait until the user confirms that the required permissions have been granted
6. Parameter Confirmation
IMPORTANT: Parameter Confirmation — Before executing any command or API call,
ALL user-customizable parameters (e.g., RegionId, instance names, CIDR blocks,
passwords, domain names, resource specifications, etc.) MUST be confirmed with the
user. Do NOT assume or use default values without explicit user approval.
Parameter
Required/Optional
Description
Default Value
Instance IP or Source IP or InstanceId
Required
Query target. Instance IP can derive InstanceId; source IP needs InstanceId (or an instance IP) as a companion
—
profile
Required
aliyun CLI profile name
The currently active profile shown by aliyun configure list
region
Optional
Region of the protection instance
cn-hangzhou
Time range
Optional
Investigation window, max 30 days
Last 1 hour
--src-ip
Optional
Filter intercept records by source IP
—
--destination-ip
Optional
Filter intercept records by destination IP
The resolved instance IP
--network-protocol
Optional
tcp / udp / icmp
—
--source-port / --destination-port
Optional
Port filters
—
--product-type
Conditional
Required only with --type default for list-policy; values: eip / ecs / slb / gf-eip
Derived from the IP's asset type, default eip
Best Practice — Two-Step Collection
Core principle: AskUserQuestion is only for choosing among known options. Never use it for free-text core identifiers like IP / InstanceId.
Step 1 — Collect core identifier in plain text.
If the user's initial message lacks an IP or InstanceId (e.g. "check recent intercepts"), respond in plain text:
Please provide the query target:
Instance IP — e.g. 47.118.170.18
Source IP + Instance IP — e.g. "source IP 1.2.3.4, instance IP 47.118.170.18"
Or an InstanceId directly — e.g. ddosbgp-cn-xxx / ddosorigin_cn-xxx
Step 2 — One AskUserQuestion for the 3 config items.
First run aliyun configure list, then send ONE AskUserQuestion with 3 questions: profile / region / time range.
Mark the active profile (the row with *) as "(active)" and list it first.
Skip Invalid profiles.
Time range options: Last 1 hour (default) / Last 24 hours / Last 7 days. The built-in "Other" field is the only place where free-text is allowed (natural-language time descriptions like "5/18 14:00 to 5/19 02:00") — the model parses them into timestamps.
7. Core Workflow
[MUST] Enable AI-Mode — AI-mode is required for Agent Skill execution.
Run the following commands before any CLI invocation:
[MUST] Disable AI-Mode at EVERY exit point — Before delivering the final response for ANY reason, always disable AI-mode first. This applies to ALL exit paths: workflow success, workflow failure, error/exception, user cancellation, session end, or any other scenario where no further CLI commands will be executed.
AI-mode is only used for Agent Skill invocation scenarios and MUST NOT remain enabled after the skill stops running.
aliyun configure ai-mode disable
Command Template
Default to plugin-mode kebab-case for every ddosbgp / antiddos-public call. Always include the User-Agent header for audit/traceability.
Edge case — OpenAPI fallback when the plugin doesn't declare an action
If a ddosbgp / antiddos-public action is missing from the installed plugin metadata, the CLI rejects it with unknown command or '<Action>' is not a valid api. In that case, bypass the plugin and go through the core CLI's built-in OpenAPI metadata:
--auto-plugin-install false → bypasses plugin routing; lets the call fall through to the core CLI's built-in OpenAPI metadata.
--force → skips parameter-validation noise on the legacy metadata path.
Use PascalCase action and parameters here (parsed by the legacy OpenAPI parser, not the plugin's kebab parser).
Do not pass --version — the core CLI uses its built-in default for the product; explicit --version is rejected as unchecked version on this route.
Known case as of aliyun-cli-ddosbgp 0.3.0: DescribeNetworkLayerIntercepts is not yet in plugin metadata, so this fallback applies. Switch back to plugin-mode kebab-case once a newer release declares describe-network-layer-intercepts.
Step 7.1 — Resolve IP role → derive InstanceId
Skip this step if the user already supplied an InstanceId.
Scenario A — IP is an Instance IP (protected destination IP):
Do NOT pass --version — the aliyun-cli-antiddos-public plugin silently ignores the --version flag on describe-bgp-pack-by-ip and always uses the CLI's built-in default (2017-05-18), which already returns the DdosbgpInfo shape this skill needs. Passing --version 2017-05-23 looks helpful but is dropped on the wire — the request still goes out as 2017-05-18. Just omit it.
Found DdosbgpInstanceId → use it as --instance-id; use the IP as --destination-ip.
Multiple DdosbgpInstanceId → list them all and ask the user which one to investigate.
⛔️ Empty DdosbgpInfo → STOP. Do NOT call any ddosbgp API. Do NOT ask the user follow-up questions.
If the response is empty (no DdosbgpInfo / DdosbgpInstanceId), the IP is not under any DDoS Native Protection pack in the current profile / region. Calling describe-network-layer-intercepts, list-policy-attachment, or list-policy next will either error or return misleading empty data that the user mistakes for "no intercepts."
Instead, output the following 4-direction checklist as a declarative final answer (statements, not questions) and END the turn. Do NOT phrase any of them as "请确认..." / "是否是占位符？" / "需要我改查 X 吗？" / "Would you like me to...?" — those create an awaiting_user_input loop. The user came here for an answer, not for an interrogation; they will re-issue the command themselves if they want a different target.
Template (translate to user's language):
已通过 DescribeBgpPackByIp 查询 IP <ip> 在 <region> / profile <active-profile>，未返回任何 DdosbgpInfo，该 IP 当前不在任何 DDoS 原生防护包内。可能的原因（按概率排序）：
IP 拼写错误 —— 请核对 <ip> 是否就是要查的目标。
profile / region 不匹配 —— 当前 profile <active-profile> / region <region> 名下不包含此 IP。如果该 IP 归属其他账号或地域，请切换 profile 或 region 后重新发起查询。
Do NOT end with "需要我帮您 X 吗？" / "请提供 Y 让我继续" / "1. ... 2. ... 请选择" — these are forbidden phrasings on this branch.
Scenario B — IP is a Source IP: source IPs can't derive an instance. The user must also supply an InstanceId or an instance IP (then follow Scenario A). Use their source IP as --src-ip later.
Scenario C — User gave NO IP and NO InstanceId (e.g., "list all my native-protection instances" / "查一下当前账号下所有实例绑定的 L3 策略"):
[MUST] --biz-region-id is required — describe-instance-list filters instances by their hosted region via --biz-region-id, not --region (the latter only routes the API call). Omitting --biz-region-id returns DDosBgp.CheckError.InvalidRegion. Pass the same region value to both flags unless you're investigating cross-region.
Returns all Anti-DDoS Origin instances under the active profile/region with their InstanceIds.
For each InstanceId, you can then proceed to Steps 7.2 / 7.4 directly — no IP derivation needed.
Do NOT skip this step and dive straight into list-policy-attachment / list-policy — without an explicit IP filter those calls either fail or return data you can't attribute back to a specific instance.
Edge case — plugin doesn't recognize the action (currently the case on aliyun-cli-ddosbgp 0.3.0): if the call above errors with unknown command or 'DescribeNetworkLayerIntercepts' is not a valid api, fall back to the OpenAPI route per the Command Template's edge-case block:
Optional source/destination filters in fallback form use PascalCase: --SrcIp, --DestinationIp, --NetworkProtocol, --SourcePort, --DestinationPort, --ProtocolNumber.
30-day cap (client-side enforcement): the API rejects windows > 30 days. If the user asks for "the last 60 days", split into two 30-day calls — do not let the server reject it.
Fallback when filtering by InterceptModule (the API has no native module filter):
Fetch page 1 (PageSize 50) and scan for the target module.
If still absent and TotalCnt > 500 → ask the user for a client source IP to filter precisely via --src-ip. If TotalCnt ≤ 500 → paginate through all records.
Step 7.3 — Analyze records and map modules
Group records by DestinationIp and InterceptModule. Use this reference (also the attribution table):
InterceptModule
Meaning
Policy field
Attribution
global_blacklist
System-level global blacklist
—
System (not user policy)
dip_blacklist
Blacklist / auto-blacklist
BlackIpList / WhiteIpList
IP-specific Mitigation Policy
packet_filter
Port blocking
PortRuleList
IP-specific Mitigation Policy or default policy (see edge case)
ipmap
Geo-block
RegionBlockCountryList / RegionBlockProvinceList
IP-specific Mitigation Policy
src_iprate
Source rate limit
SourceLimit (Bps/Pps/SynBps/SynPps)
IP-specific Mitigation Policy
l7_fp
Fingerprint filtering (payload)
FingerPrintRuleList
IP-specific Mitigation Policy
l7_filter
Fingerprint filtering (ACL)
FingerPrintRuleList
IP-specific Mitigation Policy
other
Default-policy built-in protection (incl. AI intelligent protection)
EnableIntelligence / IntelligenceLevel
Default policy
Step 7.4 — Policy correlation
List all policies bound to the IP (one call returns the default policy + IP-specific Mitigation Policy + Port-specific Mitigation Policy — do NOT pass --policy-type):
Custom policies — IP-specific Mitigation Policy (raw API PolicyType: l3) and Port-specific Mitigation Policy (raw API PolicyType: l4) — query by name only, NO --type, NO --product-type:
--product-type is mandatory with --type default. Missing it returns Invalid params. Invalid product type. — this is a parameter error, NOT an account subscription issue.
--name is server-side IGNORED on this --type default path. The API returns ALL default templates available for the given --product-type regardless of what --name you pass — passing normal, loose, or a non-existent name all return the same full template list (e.g. [eip, gf_ecs_sip_allow_flag_defense, loose, strict, normal] for --product-type eip). You MUST filter client-side: pick the entry where PolicyList[*].Name == <wanted-name>. Do NOT assume PolicyList[0] is the requested template — it usually is not. (Verified on aliyun-cli-ddosbgp 0.3.0, cn-hangzhou, 2026-05-21.)
Determining --product-type (in order of reliability): reuse a prior describe-instance-ip-address result → trial-and-error eip → ecs → slb → gf-eip (for DDoS Native Protection, eip is most common) → ask the user. Default templates fall into protection-level templates (normal / strict / loose) and scenario templates (game_tcp / game_udp / office_network / gf_origin_protect_eip_*). Map Name → console display name per references/default-policy-details.md; show only the display name to the user.
Cross-reference and attribute each record:
InterceptModule
Investigation focus
dip_blacklist
Check BlackIpList; if SourceIp not listed, it's auto-blacklist from source rate-limit (5 threshold violations in 60 s)
packet_filter
Compare PortRuleList against intercept protocol/port; if no match, see references/packet-filter-edge-cases.md
ipmap
Verify SourceIp geo against RegionBlockCountryList / RegionBlockProvinceList
This skill does NOT modify policies. When the investigation confirms a false positive, emit text guidance per the table below.
Core principle: Add the source IP to the IP-specific Mitigation Policy whitelist first. Whitelist outranks port blocking, fingerprint filtering, geo-block, source rate-limit, and reflection filtering (but is lower than blacklist).
InterceptModule
Recommendation
ActionType
Console path
other
(1) Add source IP to the IP-specific Mitigation Policy whitelist (recommended); (2) Switch default policy to a looser template
System-level — submit an Alibaba Cloud support ticket
—
Submit a ticket
Step 7.6 — Final Report (5 sections)
Intercept Overview — instance IP, InstanceId, query time window, TotalCnt, current sample count.
⛔️ Forbidden fields in this section — BaseThreshold / ElasticThreshold / ExpireTime / PackConfig.* / any billing or capacity spec. These are protection pack attributes (a "what you bought" dimension), not intercept analysis. Including them is a hard fail — the user will think this skill is also reporting pack info, which it is not. If you find yourself tempted because describe-bgp-pack-by-ip returned these fields too, drop them now.
Intercept Distribution — counts grouped by InterceptModule.
Policy Correlation — for every bound custom policy, list all modules of its type. Empty modules display "Not configured" — never silently omit.
Empty intercepts (InterceptionRecordCount = 0) — try wider time window, drop --src-ip / --network-protocol / --destination-port filters, re-verify InstanceId / region / profile. Report what was tried; never just say "no intercepts."
describe-bgp-pack-by-ip returns empty — do NOT proceed to describe-network-layer-intercepts. Verify IP, profile, region, and that the IP truly subscribes to DDoS Native Protection (vs. DDoS Pro etc.).
Time window > 30 days — split client-side; do not let the server reject it.
Forbidden.RAM / NoPermission — surface the missing Action, then point the user to references/ram-policies.md. Trigger the Permission Failure Handling block in Section 5.
Timeout / 5xx / Throttling.User — retry 1-2 times (2-5 s interval); reduce PageSize / time window; never swallow timeouts as "no intercepts" — surface the original error with RequestId.
is not a valid command or product — install the missing plugin (aliyun plugin install --names aliyun-cli-ddosbgp aliyun-cli-antiddos-public).
Exit checkpoint — Disable AI-Mode: Before producing the Final Report (success path), and at every other exit (error, cancellation, permission gate, etc.), execute aliyun configure ai-mode disable first.
8. Success Verification Method
Step-by-step verification commands and a 9-point output checklist live in references/verification-method.md. Use it to confirm the final report is complete and well-formed before delivering it.
9. Cleanup
This skill is read-only: no Alibaba Cloud resources are created, modified, or scheduled. The only state-changing operations are the AI-mode toggles in Section 7, which are explicitly disabled at every exit:
aliyun configure ai-mode disable
Nothing else needs to be torn down.
10. Command Tables
The full table of CLI commands used by this skill — including every parameter, valid enum values, JSON formats, and timestamp helpers — lives in references/related-commands.md.
11. Best Practices
Optimistic execution, diagnose on error — never preflight aliyun version / aliyun plugin list / aliyun configure list before the user has asked a concrete question (except the one targeted configure list needed to populate the profile question).
Always include --header User-Agent=... on every ddosbgp / antiddos-public call so Skills traffic can be audited.
Default to plugin-mode kebab-case for all 6 APIs. Only when the plugin rejects an action (unknown command / 'X' is not a valid api) — meaning that action is not in the installed plugin metadata — fall back to the OpenAPI route: aliyun --auto-plugin-install false <product> <PascalCaseAction> --force with PascalCase params and no--version. As of aliyun-cli-ddosbgp 0.3.0, this applies to DescribeNetworkLayerIntercepts. See Section 7 Command Template.
Two-step user-input collection — plain text for IP / InstanceId (precise format required), AskUserQuestion only for profile / region / time range. "Other" is allowed only for time range.
Whitelist-first remediation — the IP-specific Mitigation Policy whitelist beats port blocking / fingerprint / geo-block / source rate-limit / reflection filtering. Recommend it first in every false-positive scenario.
Never confuse pack thresholds with intercept analysis — BaseThreshold / ElasticThreshold are scrubbing-trigger thresholds, not rule-intercept results. Keep them out of the Intercept Overview.
List every module of both IP-specific and Port-specific Mitigation Policies in Policy Correlation, marking empty ones "Not configured" — so the user can distinguish "checked and empty" from "forgot to check."
For packet_filter records that don't match PortRuleList, attribute to the default policy (template-built-in or AI-deployed). Never claim "system anomaly" or "data mismatch." See references/packet-filter-edge-cases.md.
Surface RequestId on every unresolved error, especially timeouts and 5xx — never silently swallow as "no intercepts."
Disable AI-mode at every exit path — success, failure, error, cancellation. AI-mode must not persist after the skill stops.
⛔️ Prefer declarative output over interrogating the user — when uncertain (empty results, missing context, ambiguous request), output your best findings + suggested next steps as statements, not as questions back at the user. Avoid "请确认 X？" / "是否需要 Y？" / "Would you like me to Z?" / multiple-choice menus at the end of an answer. Only use AskUserQuestion in the initial parameter-collection phase (Section 6 Step 2), never mid-workflow or at the close of a refusal/empty/error response. The evaluator treats trailing questions as awaiting_user_input and counts the run as failed.
When the user asks about a specific default-policy template (e.g. "what does normal block?", "loose vs strict?") or when displaying a default-policy Name to the user
