edr-pro
Deploy and operate endpoint detection and response — policy tuning, behavioral detections, threat hunting, and response actions.
- License 许可
- MIT
- 最近更新
- 2026-11-07
- Stars
- ★ 0
- Forks
- 1
01 前言 Front Matter
- name
- edr-pro
- category
- security
- description
- Deploy and operate endpoint detection and response — policy tuning, behavioral detections, threat hunting, and response actions.
03 SKILL.md 正文
Overview
EDR is the defender's eyes on the endpoint: it records process, file, registry, and network activity, detects malicious behavior, and enables remote response (isolate, kill, remediate). It is arguably the highest-value single security control for most organizations — and also one of the most misconfigured, oscillating between alert floods and silent monitor-mode deployments.
This skill covers the full EDR practice: deployment, policy design, detection tuning, threat hunting on EDR telemetry, and safe response operations.
Deploy for coverage first, then tune for precision: get the agent everywhere (including servers, not just laptops), confirm telemetry is flowing, and only then start shaping policies and detections. An EDR covering 60% of endpoints with perfect tuning protects less than one covering 98% with acceptable noise.
When to use
- Rolling out EDR to endpoints and servers for the first time.
- Tuning noisy policies or investigating EDR-caused performance issues.
- Building behavioral detections for your threat model.
- Running threat hunts across endpoint telemetry.
- Defining response playbooks (isolate, contain, remediate) with approval gates.
Core concepts
-
Coverage is the metric that matters most. Unprotected endpoints are invisible endpoints. Track deployment % by asset class relentlessly — servers and OT-adjacent systems included.
-
Prevention vs detection modes. Modern EDR does both: block known-malicious behaviors automatically, detect-and-alert on suspicious ones. Progressively enable blocking on high-confidence behaviors; keep detection on the ambiguous.
-
Behavioral detections beat IOCs. Detect techniques (credential dumping, LSASS access, suspicious parent-child process chains) mapped to ATT&CK — they survive attacker infrastructure changes.
-
Policy layering. Baseline policies for all endpoints, stricter for servers and high-risk groups, monitor-only for fragile/legacy systems with compensating controls. One global policy fits nobody.
-
Telemetry retention. EDR telemetry is your best hunting and investigation dataset — retain as long as budget allows (90+ days ideal), and know exactly what you keep.
-
Response actions with gates. Network isolation, process kill, and file quarantine are powerful — define who may trigger them, require approval for broad actions, and log everything.
-
Tamper protection. Attackers kill EDR first. Enforce tamper protection, alert on agent uninstall/disable attempts, and monitor agent health as a security signal.
-
Performance partnership. EDR scans cost CPU/IO — work with IT on exclusions for legitimate heavy workloads (build servers, databases) via documented, reviewed exception processes, not ad-hoc disabling.
-
Threat hunting. Proactive hypothesis-driven searches across EDR telemetry ("show me all LSASS access outside expected processes") catch what automated detections miss.
-
Integration. EDR alerts feed the SIEM/SOC with process-tree context; EDR telemetry enriches firewall and identity investigations. An isolated EDR console is half the value.
-
Cloud workload coverage. Containers and serverless need their own sensor strategy — classic host agents do not cover ephemeral workloads. Plan EDR/CWPP coverage per workload type.
-
Offline and air-gapped endpoints. Devices off-network for weeks miss policy updates and detections. Define maximum offline tolerance and catch-up procedures.
Practical workflow
- Deploy for coverage: phased rollout (IT pilot → general population → servers → tricky segments), tracking install % daily. Resolve the long tail — the last 5% takes 50% of the effort and is often the riskiest.
- Verify telemetry: confirm events flowing per endpoint, check for blind spots (agent version skew, policy gaps, offline systems), and alert on agent health failures.
- Tune policies: start from vendor recommended baselines, adjust for your environment, document every exclusion with justification and review date. Re-tune quarterly.
- Build behavioral detections: map your top ATT&CK techniques to EDR detection rules; test against benign baselines; deploy with runbooks linked.
- Hunt regularly: weekly hypothesis-driven hunts using threat intel and red-team TTPs; convert hunt findings into permanent detections.
- Operate response: playbooks for isolate/kill/quarantine with approval gates; post-action verification (re-scan, monitor for re-infection); lessons fed back into detections and policy.
Quick wins
- Enable tamper protection everywhere this week — it is the cheapest high-value control
- Audit agent coverage by asset class; close the server gap first
- Review and justify every policy exclusion older than 90 days
Sustaining the practice
- Review policy exclusions quarterly — each needs re-justification or removal
- Re-run coverage audits monthly; new systems must get agents at provisioning
- Hunt cadence: at least weekly, tied to current threat intel
- Test tamper-protection and agent-health alerting with simulated disable attempts
Metrics that prove it works
- Agent coverage % by asset class (target: 98%+)
- Mean time to detect and to contain on endpoint incidents
- False-positive rate per policy/detection; analyst action rate
- Threat-hunt findings converted to detections per quarter
Common pitfalls
- Monitor-only forever. Detection without blocking on high-confidence behaviors leaves dwell time on the table. Progressively enable prevention.
- Coverage theater. "Deployed" with 70% coverage and no plan for the rest. The uncovered 30% is where attackers live.
- Exclusion sprawl. Every performance complaint becomes a permanent exclusion. Review exclusions quarterly with security+IT jointly.
- Killing the agent to 'fix' performance. Disabling EDR for speed is a security incident, not a fix. Tune, exclude narrowly, or upgrade hardware.
- No tamper protection. Attackers' first move is blinding the sensor. Enforce it and alert on tampering attempts.
- Alert-only hunting. EDR telemetry unused for proactive hunting wastes its best capability. Schedule hunts like any other security operation.
- Forgetting servers. Laptop-focused rollouts leave the crown-jewel servers unprotected. Servers get agents with server-tuned policies.
- Response without verification. Isolating a host and calling it done misses persistence and lateral movement. Verify with re-scan and monitoring.
- Treating EDR as antivirus replacement only. Its investigation and hunting value dwarfs the blocking value — train analysts on the telemetry, not just the alerts.
- No offline-device strategy. Laptops off VPN for a month return with stale policies and missed detections. Enforce check-in requirements.
04 同仓库更多 Skills
web-scrape
Turn web pages into LLM-ready markdown via the Firecrawl API (renders JS, anti-bot handling). Use when browser_open/fetch fails on JS-heavy pages, or when you need clean markdown/structured extraction from a site. Requires FIRECRAWL_API_KEY.
design-taste-frontend
tasteskill v2, the anti-slop frontend skill. Art direction for websites, distilled from a study of 247 curated gallery sites (One Page Love, Awwwards Sites of the Day). Deals researched directions (idea, composition, type, palette, photographs, section blueprints) from a library, enforces a strict no-slop house style with a real-browser inspector, and runs a look-and-fix loop on screenshots. Use this whenever you build, design, redesign, restyle, polish or review a landing page, marketing site, portfolio, product page, one-pager, hero or any public-facing web page, in any stack (HTML/CSS, React, Next.js, Vue, Svelte, Astro, Tailwind) — also when the user says a page looks generic, templated, "AI-made", sloppy, or asks to make it look designed, premium, award-level or less like every other site. Not for dashboards or app screens.
test-driven-development
Use when implementing any feature or bugfix, before writing implementation code
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes
requesting-code-review
Use when completing tasks, implementing major features, or before merging to verify work meets requirements
brainstorming
You MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation).