documentation-generator

Generates structured engineering documents from templates for ADRs, RFCs, runbooks, and test plans.

2|1|Updated Feb 12, 2026
One-click install
npx skills add https://github.com/cilerler/lillian --skill documentation-generator-cilerler
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: documentation-generator
Source: https://github.com/cilerler/lillian/tree/main/plugins/ai-toolkit/skills/documentation-generator
Command: npx skills add https://github.com/cilerler/lillian --skill documentation-generator-cilerler

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) and assets (resource) components.

What problem does it solve? Engineering teams lose decisions, context, and operational knowledge when documentation lacks structure. This Skill provides 20 standardized templates covering the full project lifecycle so every document type has a consistent, complete format. ## Core Features & Use Cases - Lifecycle-Organized Templates: 20 templates mapped to 10 project phases plus orthogonal incident and HR flows, including ADRs, RFCs, design docs, runbooks, SOPs, post-incident reviews, business cases, test plans, and retrospectives. - Governance Conventions: Defines identifier schemes (e.g., ADR-yyyyMMddHHmm-slug), metadata field ordering, status vocabulary, and codename-versus-functional-name rules so documents stay consistent across authors. - Use Case: When an engineer needs to record why the team chose event sourcing, the Skill directs them to read the ADR template in full, author from its exact headings, assign the timestamped ID, and place the file per the solution-structure naming rules. ## Quick Start Ask the agent to create an ADR documenting a specific architecture decision using the documentation-generator templates.

Frequently Asked Questions about documentation-generator

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

FAQPage Schema
How do I write an architecture decision record (ADR)?

Read the templates/architecture-decision-record.md file in full, then author using its exact headings: Context, Assumptions, Decision Drivers, Considered Options, Decision, Consequences, and Risks. Assign an ID in the format ADR-yyyyMMddHHmm-slug in the Metadata block, not the title.

What is the difference between an RFC, ADR, and design doc?

An RFC proposes whether to do something, an ADR immutably records one decision that was made, and a design doc describes how to build it. The default order is RFC to ADR to Design Doc, but each change picks the smallest subset that fits.

When should I create a test plan versus just test cases?

Create a test plan for multi-feature releases, regulated contexts like FDA or IEC 62304, or non-functional strategy spanning features. Skip it for feature-by-feature work where acceptance criteria, per-feature test cases, and CI gates already cover verification.

Can I update an existing ADR with new template sections?

No. Do not retrofit template sections into an already-accepted ADR by inventing content. You may restructure and rename freely, but fabricating sections like Assumptions or Risks corrupts the historical record.

Where should generated documents be placed in the repository?

Placement and naming are governed by the solution-structure skill's Documentation Placement and Naming Rules section, which is the sole source of truth. This skill does not define its own parallel filename or directory conventions.

Which documents from this skill should not be committed to the repo?

Brag documents and performance improvement plans are personal HR artifacts and must not be committed. Role briefs are intake documents handed to HR, and project intake ideas live as GitHub issues rather than repo documents.