agentleFS
PerryLinkPerryLink/dsh-skill-pack-security/skills/incident-response/SKILL.md

incident-response

Skill22 starsChanged 3 months agoApache-2.0View on GitHub

agent 环境安全事件响应:分类→控制蔓延→取证留痕→恢复→复盘的分步处置流程,覆盖密钥泄露、提示注入触发、依赖投毒、未授权操作四类事件,每条处置附命令证据。DSH/agent 环境出现疑似安全事件需要按流程处置与复盘时用;日常开发与例行维护不用。

  • Deletes or force-pushes

81 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use dora-metrics, background-jobs and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • jholhewresSkill

    dora-metrics

    Calculate DORA metrics from git + deploy logs

    Also in 3 projects that use this

  • YepAPISkill

    background-jobs

    Queue workers with BullMQ/Inngest/Trigger.dev, job retries, dead letter queues, concurrency.

    Also in 4 projects that use this

  • MicrosoftSkill

    data-modeling

    Use when the app needs to store or read data — records, a database, entities, CRUD, lists, "save/remember X", or per-user data with row-level security. This app's data layer is Rayfin's: you declare entities as TypeScript classes with decorators in rayfin/data/, register them in the schema, and read/write them through the typed rayfin-client. Covers: defining an @entity, field decorators, registering it, per-entity access with @role, owner-only row-level security, and the client query/mutation API. Triggers: data, database, model, entity, schema, table, record, CRUD, create, read, update, delete, list, store, save, persist, per-user, row-level security, RLS, access control, ownership.

    Also in 7 projects that use this

  • YepAPISkill

    data-pipelines

    ETL patterns, cron scheduling, queue-based processing, and batch jobs with Inngest/BullMQ.

    Also in 4 projects that use this

  • mmondoraSkill

    infrastructure-as-code

    Infrastructure as Code with Terraform and Pulumi. State management, modularity, drift detection, secrets handling. Use when provisioning cloud resources, writing Terraform/Pulumi modules, or managing infrastructure configuration.

    Also in 5 projects that use this

  • csiddhant796-blipSkill

    threat-detection

    Use when hunting for threats in an environment, analyzing IOCs, or detecting behavioral anomalies in telemetry. Covers hypothesis-driven threat hunting, IOC sweep generation, z-score anomaly detection, and MITRE ATT&CK-mapped signal prioritization.

    Also in 3 projects that use this

The file

Files
SKILL.md

事件响应(incident-response)

原则:先控制,再取证,后恢复。控制动作(轮换/撤销/停用)不依赖完整归因——疑似即执行;取证在控制之后做,避免"边查边扩散"。全程不写密钥明文,报告遵守 secret-scan 脱敏规范。

1. 确认与分类

先回答两个问题:发生了什么(类型),影响面多大(范围)。四类事件与初步证据:

事件类型初步证据立即动作(见第 2 节)
密钥泄露扫描告警、公开仓库出现 token轮换 + 撤销
提示注入触发会话执行了上下文外指令(下载/外发/改配置)停用触发源 + 隔离
依赖投毒新依赖出现异常 install 脚本/网络请求回滚依赖 + 锁文件冻结
未授权操作仓库/CI 出现非本人提交、非预期运行撤销凭据 + 暂停自动化
git rev-parse HEAD
git log -1 --format='%H %cd' --date=iso-strict

预期输出样例:a1b2c3d4... 与提交时间。 判据:先记录当前 HEAD——后续所有证据命令都相对它复核;不确定类型时先按"密钥泄露"假设处置(轮换成本最低、最安全)。

2. 控制蔓延(疑似即执行,不等归因)

  • 密钥类:按 secret-scan 第 6 节顺序——先轮换、再撤销;轮换完成判据 = 新值生效且旧值在厂商控制台已撤销。
  • 注入/未授权操作类:停用触发源——下线相关 MCP/插件、暂停 agent 会话与相关 CI:
git ls-files -- 'cordis.yml' '**/cordis.yml' '.mcp.json' '.github/workflows/**'
dsh --profile <配置文件> --dump-config

样例输出:配置与工作流路径列表;dump 输出中能看到已挂载插件清单。 判据:dump 输出里逐个核对插件/MCP 是否受信;存疑的立即从配置移除(改配置 → 新会话生效)。控制完成判据 = 事件源已停用且 24 小时内无同类新动作。

3. 取证留痕(时间线 + 证据包)

git log --format='%H %cd %s' --date=iso-strict --since='<事发时间>' --all
# Windows:Get-ChildItem $env:DSH_HOME -Recurse -File | Sort-Object LastWriteTime -Descending | Select-Object -First 20
# macOS/Linux:find ~/.dsh -type f -newermt '<事发时间>' -ls

样例输出:一行一条提交(哈希 + 时间 + 摘要);会话目录里按时间倒序的文件列表。 判据:时间线每条 = 时间 + 主体 + 动作 + 证据引用;只记录可复核的事实,推测写进"待验证"区。证据包 = 命令输出原文 + 相关文件副本,按 secret-scan 脱敏规范处理后再共享。

4. 恢复

git revert --no-commit <可疑提交>
# 或回退整个分支到已知良好提交:git reset --hard <已知良好提交>(先备份!)

样例输出:revert 后 git status 显示待提交的逆变更。 判据:恢复顺序——先撤恶意/可疑变更,再恢复受影响服务,最后恢复自动化(CI/agent)并观察一个周期。恢复完成判据 = 时间线中所有异常点都有对应处置记录,且观察期内无复发。

5. 复盘与加固

复盘模板(五段:时间线、根因、影响评估、处置记录、加固项)见 references/runbook-and-postmortem.md。加固项从根因反推,每项都要可验证:

  • 密钥类根因 → 补 gitleaks protect --staged 门禁(secret-scan);
  • 注入类根因 → 按 prompt-injection-review 过一遍注入面并落地缓解;
  • 依赖类根因 → 按 supply-chain-review 收紧引入流程(pin action、禁 install 脚本);
  • 权限类根因 → 最小权限改造:删除多余凭据、限制 CI 令牌范围。

复盘完成判据 = 每个加固项都有验证命令或验收标准。

与其他技能的分工

  • secret-scan:密钥泄露事件的检测、分级与轮换流程细节。
  • prompt-injection-review:注入类事件的注入面排查与缓解清单。
  • supply-chain-review / dependency-audit:依赖类事件的引入向量分析与依赖审计。
  • security-audit:事件平息后的全量审计复查。

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.