ai-architect
Scannednpx machina-cli add skill cloudaipro/openclaw-agent-skills/ai-architect --openclawAI Architect
Objective
Design robust, implementation-ready skill specs that can be scaffolded into Codex/OpenClaw skill folders.
Trigger Rules
Use when users request new skill design or skill-system architecture.
Positive cues:
- "Create a new skill for ..."
- "Draft SKILL.md for this role."
- "Design an agent workflow and folder structure."
Do not use when:
- User requests only explanation of external trends.
- User requests only root-cause social analysis.
- User asks for operational planning without skill design.
Inputs
Required:
- Skill purpose and core task.
Optional:
- Tooling constraints.
- Permission profile.
- Output schema requirements.
Output Schema
Return universal envelope from ../shared/references/output-schemas.md.
Artifacts order:
- Skill architecture brief.
- Trigger and resource matrix.
- Ready-to-edit SKILL.md draft and scaffold plan.
- Validation and rollout checklist.
Workflow
- Clarify scope and trigger boundaries.
- Define input/output contract and permission model.
- Propose folder layout with scripts/references/assets as needed.
- Draft SKILL.md and runtime implementation notes.
- Define validation gates and release phases.
Quality Bar
- Trigger boundaries must be explicit and testable.
- Output contract must be schema-aligned.
- Rollout must include evaluation and rollback logic.
Safety Rules
- Keep least-privilege defaults.
- Require explicit approval for write-side actions.
- Avoid ambiguous or non-verifiable requirements.
Resources
- Domain framework:
references/domain.md - Envelope validator:
scripts/validate_output.py - Reasoning artifact validator:
scripts/validate_reasoning_artifacts.py
Source
git clone https://github.com/cloudaipro/openclaw-agent-skills/blob/main/skills/ai-architect/SKILL.mdView on GitHub Overview
ai-architect designs robust, implementation-ready skill specs that can be scaffolded into Codex/OpenClaw folders. It drafts SKILL.md specifications, trigger boundaries, resources, and validation gates to support creation or refactor of AI skills, role prompts, or agent workflows.
How This Skill Works
Follows a structured workflow: clarify scope and trigger boundaries, define input/output contracts and permission models, propose a folder layout with scripts/references/assets, draft SKILL.md and runtime notes, and define validation gates plus rollout plans. Outputs are aligned to a universal envelope schema and delivered as a ready-to-edit blueprint.
When to Use It
- Designing a brand-new skill for a specific task and workflow.
- Drafting SKILL.md for a new or updated role prompt or agent capability.
- Designing an agent workflow and folder structure with scripts, references, and assets.
- Refactoring an existing AI skill to tighten trigger boundaries, IO contracts, and permissions.
- Defining trigger boundaries and validation gates for rollout and rollback.
Quick Start
- Step 1: Clarify scope and trigger boundaries with the user request.
- Step 2: Define input/output contracts and permission model; sketch folder layout.
- Step 3: Draft SKILL.md, add runtime notes, and outline validation gates and rollout plan.
Best Practices
- Define explicit, testable trigger boundaries before drafting any SKILL.md.
- Anchor all outputs to the shared IO contract and output schema.
- Propose a clear folder layout including scripts/references/assets.
- Include rollout, evaluation, and rollback plans in the blueprint.
- Enforce least-privilege defaults and require explicit approvals for write actions.
Example Use Cases
- Design a new CustomerSupportResponder skill with a defined IO contract and permission model.
- Draft SKILL.md for a DataCollectionAgent role prompt with clear responsibilities.
- Create a folder scaffold for OrderFulfillment that maps scripts, references, and assets.
- Refactor a SentimentAnalysis skill to tighten trigger boundaries and add validation gates.
- Plan a rollout for a ComplianceCheck skill including evaluation metrics and rollback steps.
