chaos-and-degradation-choreographer
Skill4 starsChanged 31 days agoMITView on GitHub
Plan and review graceful degradation and controlled fault tests for dependencies, worker pools, retries, circuit breakers, and recovery. Use when an optional dependency may fail, a critical outcome becomes uncertain, or recovery could overload a service.
In this repository47 other files it gives its agents.
- AGENTS.mdAGENTS.md
- anti-slop-engineeringSkill
- anti-slop-engineeringSkill
- api-ergonomics-and-contract-governorSkill
- apple-craft-and-frame-budgetSkill
- apple-craft-and-interaction-physicsSkill
- asset-provenance-guardSkill
- backprop-guardSkill
- chromatic-drift-guardSkill
- cognitive-load-and-attention-governorSkill
- comment-budget-guardSkill
- context-window-and-token-economics-architectSkill
- cryptographic-agility-and-secret-lifecycle-gateSkill
- decision-guardSkill
- demock-guardSkill
- design-system-drift-guardSkill
- deterministic-simulation-and-property-testingSkill
- distributed-consensus-and-concurrency-gateSkill
- finops-and-unit-economics-governorSkill
- future-scale-gateSkill
- grumpy-senior-staff-engineerSkill
- honeypot-trap-guardSkill
- house-tell-guardSkill
- install-anti-slopSkill
- llm-eval-and-prompt-injection-fortressSkill
- mandela-guardSkill
- mechanical-sympathy-and-perf-gateSkill
- mechanical-sympathy-profilerSkill
- motion-runtime-guardSkill
- observability-and-slo-architectSkill
- observability-and-slo-gateSkill
- paranoid-principal-architectSkill
- pr-boundary-guardSkill
- privacy-and-data-lineage-gateSkill
- re0-handoff-guardSkill
- reduction-guardSkill
- repo-memory-and-invariant-stewardSkill
- root-fix-gateSkill
- safe-evolution-and-rollout-choreographerSkill
- scientific-fault-isolationSkill
- styleseed-guardSkill
- supply-chain-and-sbom-sentinelSkill
- tagore-guardSkill
- type-level-and-ast-slop-gateSkill
- ui-and-copy-disciplineSkill
- verification-oracle-architectSkill
- zero-downtime-migration-gateSkill
The file
Files
Chaos and Degradation Choreographer
Plan by dependency
- Trace required and optional dependencies per operation. Record the customer-visible result for healthy, degraded, uncertain, overloaded, and recovered states. Mark irreversible effects and authorization checks as required.
- Allocate an end-to-end deadline across calls, queueing, and retries. Give dependencies separate bounded capacity where one must not starve another. Specify when to reject excess work, who owns retries, and the total attempt budget. Classify errors before retrying. Release resources on cancellation and terminal failures.
- Write the fallback's data value, age, authorization, validation, and observability contract. A nonessential cache may degrade to bounded origin reads. An uncertain payment stays pending until reconciled, using the same provider idempotency key only if its contract permits replay. Apply the same identity and data-validation checks to the fallback.
- Define the circuit breaker's trigger, open behavior, half-open probe budget, and recovery condition. Design load shedding and recovery to avoid a surge from every queued client when the dependency becomes healthy.
- Run the normal case first and change one fault at a time in an isolated test environment with fake credentials, bounded load, controlled clock, and disabled production access. Define the stop signal, deadline, and cleanup in advance. Record affected operations, baseline metrics, fault timing, fallback use, queue recovery, and the exact assertions and command.
Schedules
| Fault | Evidence |
| Optional Redis cache unavailable | Unaffected path responds within the degraded budget. Origin traffic stays within its reserved capacity. Extra requests are shed after the cap. |
| Payment timeout after request submission | Return a pending state. Reconcile provider state before treating the attempt as definitively failed. At most one effect appears across retries. |
| Persistent transient errors | Requests end at the total attempt or time budget. Unrelated queues retain capacity. No dependency requires an unbounded backlog to recover. |
| Fallback data fails identity, freshness, or schema check | Withhold the fallback. Preserve a distinct invalid-data signal and avoid looking like a successful fresh response. |
| Dependency restored | Permit bounded probes, re-open on a failed probe, and restore traffic gradually on successful probes. Verify the queue drains and user-facing latency returns to its agreed interval. |
The offline schedule fixture exercises the decisions above using injected states and an integer budget. It runs without network, subprocesses, or real services. The fixture is a policy model. Verify latency, cancellation, failure isolation, and recovery on the real component in a disposable integration environment.
Catalog: CON-09, CON-11, OPS-07, SRE-08. See also deterministic simulation with the optional deterministic-simulation-and-property-testing skill.
References: Google SRE on cascading failures, AWS on timeouts, retries, and jitter.
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
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.

