PR Custody

Add Shift-Left Reliability skill

@rsionnachchecks n/achecks…add-shift-left-reliability → master2 files · +118 −0updated 4mo ago

▸Description

What problem it solves

Reliability requirements are typically scattered across wikis, runbooks, and tribal knowledge. This skill brings reliability decisions into the development workflow by generating OpenSRM manifests — machine-readable YAML files that declare SLOs, dependencies, ownership, observability, and deployment gates.

Who uses this workflow

SREs, platform engineers, and backend developers who want to define reliability requirements as code alongside their services, rather than discovering gaps after incidents.

What the skill does

Generates service.reliability.yaml manifests for api, worker, stream, ai-gate, batch, and database services
Suggests SLO targets based on service tier (critical/high/standard/low)
Validates that SLO targets are achievable given dependency chains
Adds judgment SLOs (reversal rate, calibration, etc.) for AI decision services
Checks that internal targets are tighter than external contract promises
Covers observability requirements and deployment gates

Example

Ask Claude: "Create a reliability manifest for my payment API service" — and it generates a spec-compliant service.reliability.yaml with appropriate SLOs, ownership, and dependency declarations.

Attribution

Based on the OpenSRM specification (Open Service Reliability Manifest).

loading diff…