chaos-engineering — Claude Code plugin: experiments and postmortems
Runs a chaos experiment end to end: hypothesis, steady-state metric, blast radius, abort criteria, and a blameless postmortem.
/plugin marketplace add alirezarezvani/claude-skills
/plugin install chaos-engineering@claude-code-skillsWhat chaos-engineering does
The plugin walks you through an experiment from design to write-up. You state a hypothesis, pick the metric that defines normal behaviour, set how much traffic takes the hit, and name the condition that stops the run. Then it works out the risk: how many users the fault reaches and how that sits against your error budget. Once the run is over, it produces a postmortem with no blame attached.
Three scripts ship inside, all on the Python standard library: experiment_designer.py, blast_radius_calculator.py, experiment_postmortem.py. Four reference docs cover the principles, experiment design, a taxonomy of 7 attack types, and the tooling field (Chaos Toolkit, Chaos Mesh, Litmus, Gremlin, AWS FIS, homegrown setups). Templates for the plan and the postmortem are included, and the /chaos-experiment slash command runs the design as an interactive wizard.
Abort criteria have to be written before anything starts. Skip that step and the experiment turns into the outage you were trying to learn about. The plugin pairs with feature-flags-architect, where kill switches double as abort triggers, and with kubernetes-operator, since operators are frequent chaos targets.
What it helps with
- design an experiment with a falsifiable hypothesis and a steady-state metric
- work out the blast radius and check it against the error budget
- write abort criteria before the run instead of during an incident
- turn a result log into a blameless postmortem
- pick one of seven attack types and set the magnitude
Example prompts for the agent
- Design a chaos experiment for
checkout-svc: hypothesis p99 < 500ms when payment slows, latency attack +200ms, abort if p99 > 1000ms - Calculate the blast radius for 5% traffic share, 1000000 users, 15 minutes
- Generate a postmortem from
plan.jsonandresults.txt
How to install chaos-engineering
Plugin marketplace
/plugin marketplace add alirezarezvani/claude-skills/plugin install chaos-engineering@claude-code-skillsStep by step, with commands for each platform: Claude Code plugins: marketplaces, installing, and building your own
Questions about chaos-engineering
How do I install chaos-engineering in Claude Code?
Through the marketplace: run /plugin marketplace add alirezarezvani/claude-skills first, then /plugin install chaos-engineering@claude-code-skills.
What does chaos-engineering do?
It covers the whole chaos experiment cycle: design with hypothesis and abort criteria, blast-radius calculation, and postmortem generation. Three Python scripts, four reference docs and two templates come with it.
Does it work alongside other skills?
Yes. With feature-flags-architect the kill switches act as abort triggers, and with kubernetes-operator the operators themselves become chaos targets.
Source text
chaos-engineering · MIT
# Chaos Engineering
End-to-end discipline for chaos experiments — design, run, learn — without becoming an outage.
## What's inside
- **3 stdlib Python tools** — experiment designer, blast-radius calculator, postmortem generator
- **4 reference docs** — principles, experiment design, attack taxonomy, tooling landscape
- **2 templates** — experiment plan, postmortem
- **`/chaos-experiment` slash command** — interactive design wizard
## Install
```bash
# Via Claude Code marketplace
/plugin install chaos-engineering
# Or clone the repo
git clone https://github.com/alirezarezvani/claude-skills.git
cd claude-skills/engineering/chaos-engineering
```
## Quick start
```bash
SKILL=engineering/chaos-engineering/skills/chaos-engineering
# 1. Design an experiment
python "$SKILL/scripts/experiment_designer.py" \
--target checkout-svc \
--hypothesis "p99 < 500ms when payment slows" \
--attack latency --magnitude "+200ms" \
--abort-if "p99 > 1000ms OR error_rate > +1pp"
# 2. Calculate blast radius
python "$SKILL/scripts/blast_radius_calculator.py" \
--traffic-share 0.05 --user-pop 1000000 \
--duration-min 15
# 3. Generate postmortem after running
python "$SKILL/scripts/experiment_postmortem.py" \
--plan plan.json --result-log results.txt
```
## Key principles
1. **Build a hypothesis around steady-state behavior** — measurable, falsifiable
2. **Vary real-world events** — realistic failures only, not astronomy-grade
3. **Run experiments in production** — staging never has prod failure modes
4. **Automate experiments to run continuously** — single experiment = press release; continuous = engineering
5. **Define abort criteria up front** — chaos without abort = outage
## The 7 attack types
| Attack | Tests | Magnitude examples |
|---|---|---|
| **Latency** | Timeouts, retries, circuit breakers | +200ms, +2s |
| **Error** | Error handling, fallback paths | 1%, 50%, 100% |
| **Resource** | Saturation, autoscaling, OOM | 80% CPU, 90% memory, fill /var |
| **Network partition** | Consensus, leader election, failover | drop 100% to peer X |
| **Dependency failure** | Graceful degradation | timeout 100% to dep X |
| **Time skew** | Clocks, TTLs, retry backoff | +5min, +1day |
| **Infrastructure** | Auto-recovery, replica maintenance | kill 1 of N |
## Composition with other skills
| Skill | Composition |
|---|---|
| `feature-flags-architect` | Kill switches there are abort triggers here |
| `kubernetes-operator` | Operators are common chaos targets |
| `incident-response` | Chaos that escalates becomes an incident |
## Skill structure
```
chaos-engineering/
├── README.md
├── .claude-plugin/plugin.json
└── skills/chaos-engineering/
├── SKILL.md
├── scripts/
│ ├── experiment_designer.py
│ ├── blast_radius_calculator.py
│ └── experiment_postmortem.py
├── references/
│ ├── chaos_principles.md
│ ├── experiment_design.md
│ ├── attack_taxonomy.md
│ └── tooling_landscape.md
└── assets/
├── experiment_template.md
└── postmortem_template.md
```
## Verifiable success
A team using this skill should achieve:
- 100% of chaos experiments have written hypothesis, abort criteria, blast-radius calc
- Blast radius for any single experiment ≤10% of monthly error budget
- Mean time between chaos experiments <14 days (continuous, not one-off)
- Each experiment produces ≥1 follow-up action that gets shipped
- No chaos experiment escalates to a customer-impacting incident in trailing 90 days
## License
MIT — see repo root LICENSE.
© Alireza Rezvani · MIT · alirezarezvani/claude-skills
