Contract
- Input: service inventory, SLOs, dependency graph, production traffic patterns.
- Output: chaos experiment plan + fault injection config + resilience report.
- Scope: designs experiments and documents results; does not execute production chaos without explicit authorisation and safety review.
- Rule: designs experiments and documents results; does not execute production chaos without explicit authorisation and safety review.
- Rule: documented standards override defaults; explicit project rules take precedence.
- Rule: if blocked by missing context or dependencies, surface the blocker before proceeding.
Provenance
| Question |
Answer |
| What is the source of truth? |
The user's request, originating spec/issue, and the skill's declared outputs |
| What is in scope? |
Work covered by the skill's acceptance criteria and completion rules |
| What is explicitly out of scope? |
Files, behaviors, and decisions outside the skill's declared boundary |
| Who or what consumes this artifact afterward? |
The next skill in the workflow or the user |
| What evidence proves it is done? |
Completion criteria met, artifact saved, validation passed |
| What risk remains? |
Subjective judgment calls, missing context, or external dependency failures |
Artifact
Emit ChaosEngineeringArtifact as both:
- JSON:
.agents/skills/platform/artifacts/chaos-engineering/{identifier}.json
- Markdown view: same filename with
.md extension
Chaos Engineering
Design chaos engineering — fault injection, GameDays, resilience validation — with explicit steady-state hypotheses, blast radius controls, and safety rails.
Process
1. Define steady state
- System behaviour under normal conditions (SLOs: latency, error rate, throughput).
- Observable metrics and dashboards for live monitoring during experiments.
- Blast radius definition (region, cluster, service, customer segment).
- Safety rails (abort conditions, automatic rollback, circuit breakers).
Completion criterion: steady-state metrics, blast radius, and safety rails documented.
2. Identify weaknesses
- Dependency map (upstream services, databases, caches, DNS, load balancers).
- Historical failure modes (outages, degradation, cascading failures).
- Single points of failure and lack of redundancy.
- Hypothesis: "If we inject fault X, the system will behave Y within SLO."
Completion criterion: weakness list and experiment hypotheses documented.
3. Design experiments
- Fault types (network latency, packet loss, DNS failure, CPU / memory pressure, instance termination, dependency failure).
- Experiment scope (production-like staging vs production; time window).
- Success and failure criteria (SLO held? system degraded? data lost?).
- Rollback plan (manual and automated).
Completion criterion: experiment design with success/failure criteria defined.
4. Run GameDay
- Pre-flight checklist (approvals, monitoring ready, on-call briefed).
- Execute fault injection with real-time observation.
- Abort immediately if safety rails are breached.
- Record timeline, observations, and system response.
Completion criterion: GameDay execution log and observation notes saved.
5. Analyze and document
- Compare observed behaviour against steady-state hypothesis.
- Identify surprising behaviours, hidden dependencies, and resilience gaps.
- Assess impact on user experience and downstream services.
- Prioritise remediation (P0 fix, P1 follow-up, accepted risk).
Completion criterion: resilience report with findings and remediation backlog defined.
6. Continuous improvement
- Automate repeatable experiments in CI/CD or staging.
- Track remediation items and verify fixes with follow-up experiments.
- Expand experiment coverage to new services and failure modes.
- Share findings across teams to build organisational resilience knowledge.
Completion criterion: automation plan and sharing process defined.
Rules
- Rule: start with a steady-state hypothesis; never inject faults without an expected outcome.
- Rule: limit blast radius; prefer staging over production for first experiments.
- Rule: require live monitoring and on-call readiness before any fault injection.
- Rule: stop immediately if abort conditions are met; do not push through to prove a point.
- Rule: treat every finding as a product risk item, not a one-off engineering curiosity.
Completion
- the skill's completion criteria are explicitly checked
- the artifact is saved and validated
- any blockers or skipped validations are documented
- the next consumer is identified or the work is handed off
@include .agents/skills/platform/contract-base.xml
1---2name: "chaos-engineering"3category: "devops"4maturity: "stable"5version: "1"6description: "Design chaos engineering — fault injection, GameDays, resilience validation — with explicit steady-state hypotheses, blast radius controls, and safety rails."7license: "MIT"8capabilities:9 - apply chaos engineering methods within explicit devops boundaries10 - analyze design chaos engineering — fault injection, gamedays, resilience validation — with explicit steady-state hypotheses, blast radius controls, and safety rails with assumptions, constraints, and failure modes11 - produce and validate a devops artifact with evidence, decisions, and next steps12outputs: "chaos experiment plan + fault injection config + resilience report."13sideEffects: []14dependencies: []15stopCondition: "Experiment plan defined; fault injection config documented; resilience report produced."16risk: "medium"17trustTier: "3"18maxIterations: "6"19promptVersion: "2.0"20artifactType: "devops"21modelTier: "reasoning"22evaluators: ["behavioral", "regression"]23fixturesPath: .agents/skills/platform/fixtures/behavioral/chaos-engineering.json24diataxis: "how-to"25tags: ["devops"]26compatibility: []27approvalRequired: "true"28approvalFor: []29metadata:30 source: https://skillmd.com/skills/quantumquirkxyz/chaos-engineering31---3233## Contract3435- Input: service inventory, SLOs, dependency graph, production traffic patterns.36- Output: chaos experiment plan + fault injection config + resilience report.37- Scope: designs experiments and documents results; does not execute production chaos without explicit authorisation and safety review.38- Rule: designs experiments and documents results; does not execute production chaos without explicit authorisation and safety review.39- Rule: documented standards override defaults; explicit project rules take precedence.40- Rule: if blocked by missing context or dependencies, surface the blocker before proceeding.4142## Provenance4344| Question | Answer |45|---|---|46| What is the source of truth? | The user's request, originating spec/issue, and the skill's declared outputs |47| What is in scope? | Work covered by the skill's acceptance criteria and completion rules |48| What is explicitly out of scope? | Files, behaviors, and decisions outside the skill's declared boundary |49| Who or what consumes this artifact afterward? | The next skill in the workflow or the user |50| What evidence proves it is done? | Completion criteria met, artifact saved, validation passed |51| What risk remains? | Subjective judgment calls, missing context, or external dependency failures525354## Artifact5556Emit `ChaosEngineeringArtifact` as both:57- JSON: `.agents/skills/platform/artifacts/chaos-engineering/{identifier}.json`58- Markdown view: same filename with `.md` extension596061# 62# 63# Chaos Engineering6465Design **chaos engineering** — fault injection, GameDays, resilience validation — with explicit steady-state hypotheses, blast radius controls, and safety rails.6667## Process6869### 1. Define steady state70- System behaviour under normal conditions (SLOs: latency, error rate, throughput).71- Observable metrics and dashboards for live monitoring during experiments.72- Blast radius definition (region, cluster, service, customer segment).73- Safety rails (abort conditions, automatic rollback, circuit breakers).7475**Completion criterion:** steady-state metrics, blast radius, and safety rails documented.7677### 2. Identify weaknesses78- Dependency map (upstream services, databases, caches, DNS, load balancers).79- Historical failure modes (outages, degradation, cascading failures).80- Single points of failure and lack of redundancy.81- Hypothesis: "If we inject fault X, the system will behave Y within SLO."8283**Completion criterion:** weakness list and experiment hypotheses documented.8485### 3. Design experiments86- Fault types (network latency, packet loss, DNS failure, CPU / memory pressure, instance termination, dependency failure).87- Experiment scope (production-like staging vs production; time window).88- Success and failure criteria (SLO held? system degraded? data lost?).89- Rollback plan (manual and automated).9091**Completion criterion:** experiment design with success/failure criteria defined.9293### 4. Run GameDay94- Pre-flight checklist (approvals, monitoring ready, on-call briefed).95- Execute fault injection with real-time observation.96- Abort immediately if safety rails are breached.97- Record timeline, observations, and system response.9899**Completion criterion:** GameDay execution log and observation notes saved.100101### 5. Analyze and document102- Compare observed behaviour against steady-state hypothesis.103- Identify surprising behaviours, hidden dependencies, and resilience gaps.104- Assess impact on user experience and downstream services.105- Prioritise remediation (P0 fix, P1 follow-up, accepted risk).106107**Completion criterion:** resilience report with findings and remediation backlog defined.108109### 6. Continuous improvement110- Automate repeatable experiments in CI/CD or staging.111- Track remediation items and verify fixes with follow-up experiments.112- Expand experiment coverage to new services and failure modes.113- Share findings across teams to build organisational resilience knowledge.114115**Completion criterion:** automation plan and sharing process defined.116117## Rules118119- Rule: start with a steady-state hypothesis; never inject faults without an expected outcome.120- Rule: limit blast radius; prefer staging over production for first experiments.121- Rule: require live monitoring and on-call readiness before any fault injection.122- Rule: stop immediately if abort conditions are met; do not push through to prove a point.123- Rule: treat every finding as a product risk item, not a one-off engineering curiosity.124125## Completion126127- the skill's completion criteria are explicitly checked128- the artifact is saved and validated129- any blockers or skipped validations are documented130- the next consumer is identified or the work is handed off131---132@include .agents/skills/platform/contract-base.xml