suede-ops-architecture

Designs operations architecture with entity schemas, write paths, and work triage before builds.

123|10|Updated May 24, 2026
One-click install
npx skills add https://github.com/JasonColapietro/suede-creator-skills --skill suede-ops-architecture-jasoncolapietro
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: suede-ops-architecture
Source: https://github.com/JasonColapietro/suede-creator-skills/tree/main/skills/suede-ops-architecture
Command: npx skills add https://github.com/JasonColapietro/suede-creator-skills --skill suede-ops-architecture-jasoncolapietro

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Teams build automations and AI agents on fragmented data and sprawling tool stacks, producing confident outputs built on partial context. This Skill fixes the shape of a system before anyone builds it, so nothing lands on a moving foundation. ## Core Features & Use Cases - Schema and write-path design: Names every entity, its fields, lifecycle states, and enforces exactly one canonical write path per entity. - Work triage: Routes every unit of work to a deterministic automation, an extraction step, an AI agent, or a human decision using the Written Rule Test, with a blast-radius override for irreversible actions. - Tool consolidation and migration planning: Marks every existing tool absorb, keep, or kill with a cited proof path, then chooses a full, hybrid, or cutover-date migration path and sequences the build into command-gated phases. - Use Case: A team with 22 tools and a spreadsheet that actually runs the business asks what to cut; the Skill requires a proof path per tool, flags the spreadsheet as a competing writer, and returns a phased architecture plan. ## Quick Start Use the suede-ops-architecture skill to design the architecture for our intake automation before we build anything.

Frequently Asked Questions about suede-ops-architecture

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I decide between automation and an AI agent for a workflow?

Apply the Written Rule Test: if the complete rule can be written down with every branch and inputs arrive structured, build deterministic code with no model. Use an agent only when the rule resists writing but competent people agree on the answer; route disagreement to a human decision.

How do I consolidate a sprawling tool stack before building a new system?

Mark each tool absorb, keep, or kill based on a cited proof path such as a last-write timestamp, seat count, or invoice line. Tools with no evidence are reported as unknown, never killed, and regulated or financial systems of record are kept and integrated.

When should an AI agent not act autonomously on workflow output?

Apply the blast-radius override: any work whose wrong output moves money, reaches people outside the team, changes access, or deletes data ships as agent-proposes and human-approves. Reversibility decides, not task difficulty.

How do I choose a data migration path off a legacy system?

Choose full migration when history is read daily, hybrid when reference records outlive the work, and a cutover date when cycle length is short and closed records are rarely reopened. Retention requirements alone are satisfied by a read-only archive, not a full migration.

What are the limitations of this operations architecture skill?

It produces the written architecture only and does not build systems, migrate data, or cancel subscriptions. Lead scoring and CRM routing belong to suede-revops, multi-file repo changes to suede-graph-flo-xr, and eval-set validation to suede-ai-eval.