Where to Find Reference Examples for the narrow-react-prop-types Skill

The narrow-react-prop-types skill provides a dedicated references/ folder containing response templates, agent memory configurations, and GitHub Actions workflow definitions that constitute the complete canonical examples for the skill.

The narrow-react-prop-types skill in the humanlayer/skills repository automates the process of narrowing React component prop types in TypeScript monorepos. Located within the skill's directory structure, the reference examples demonstrate exactly how the CodeLayer CI agent generates pull requests, persists memory across runs, and wires together automation workflows. These files serve as the source of truth for anyone customizing or extending the skill's behavior.

Location of Reference Examples in the Repository

The complete set of reference examples resides in plugins/narrow-react-prop-types/skills/narrow-react-prop-types/references/. This directory contains three critical files that illustrate the expected input format, memory persistence strategy, and end-to-end workflow orchestration.

Response Template (response-template.md)

The file references/response-template.md defines the markdown skeleton that the CI agent uses to generate pull request bodies. It structures the output summary, changes table, and rationale columns that appear in automated PRs.


## React Component Prop Narrowing Complete

### Summary

Narrowed props in **N** components.

### Changes Made

| Component | File | Change | Rationale |
|-----------|------|--------|-----------|
| `ComponentName` | `path/to/file.tsx` | Made `onAction` required | All live call sites provide it; menu item would be inert without handler |

Agent Memory (narrow-component-props-memory.md)

The references/narrow-component-props-memory.md file contains a YAML-style memory snippet that persists across skill invocations. This file guides the agent on which monorepo directories to target and establishes standing feedback for future runs.


# Agent Memory: Narrow React Component Props

Standing feedback for future `Agent: Narrow React Component Props` runs.

## Guidance

- Focus only on `apps/riptide-ui`; do not make changes in `apps/riptide-cloud`.

Workflow Definition (agent-narrow-component-props.yml)

The file references/agent-narrow-component-props.yml provides the GitHub Actions workflow that invokes the skill via CodeLayer. It wires together the memory file and response template, defining how the automation triggers on workflow_dispatch events.

name: narrow-react-prop-types
on:
  workflow_dispatch:
jobs:
  narrow-prop-types:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v3
      - name: Run CodeLayer agent
        uses: code-layer/agent@v1
        with:
          skill: narrow-react-prop-types
          memory-file: references/narrow-component-props-memory.md
          response-template: references/response-template.md

How the References Connect to SKILL.md

According to the source code in the humanlayer/skills repository, the main skill definition at plugins/narrow-react-prop-types/skills/narrow-react-prop-types/SKILL.md references these three files directly. The SKILL.md file serves as the human-readable entry point, while the references/ directory contains the machine-executable assets.

When the CodeLayer agent executes the narrow-react-prop-types skill, it reads the workflow definition to determine job configuration, loads the memory file to understand repository-specific constraints, and renders the response template to format the final PR description. This separation of concerns allows developers to modify output formatting or target scopes without changing the core skill logic.

Summary

  • The canonical reference examples for the narrow-react-prop-types skill live in plugins/narrow-react-prop-types/skills/narrow-react-prop-types/references/.
  • response-template.md controls the markdown structure of generated pull request bodies.
  • narrow-component-props-memory.md provides persistent YAML-based guidance for targeting specific monorepo directories.
  • agent-narrow-component-props.yml defines the GitHub Actions workflow that orchestrates the CodeLayer agent invocation.
  • The parent SKILL.md file links to these resources, establishing the complete reference architecture.

Frequently Asked Questions

What is the purpose of the response-template.md file?

The response-template.md file serves as the markdown blueprint for pull request descriptions generated by the narrow-react-prop-types skill. It standardizes the presentation of changes, ensuring every automated PR includes a summary count, a detailed changes table with rationale columns, and consistent formatting that human reviewers can quickly scan.

How does the agent memory file persist across runs?

The narrow-component-props-memory.md file acts as persistent storage for the CodeLayer agent, maintaining YAML-formatted guidance between workflow invocations. According to the repository structure, the workflow definition explicitly passes this file path to the agent via the memory-file parameter, allowing the skill to remember repository-specific constraints like targeting only apps/riptide-ui while excluding apps/riptide-cloud.

Where is the workflow definition located?

You can find the workflow definition at plugins/narrow-react-prop-types/skills/narrow-react-prop-types/references/agent-narrow-component-props.yml. This file configures the GitHub Actions job that runs on ubuntu-latest, checks out the repository, and invokes the CodeLayer agent with specific pointers to the memory file and response template.

Can I customize these reference files for my own repository?

Yes, the reference files are designed to be customized. You can modify narrow-component-props-memory.md to specify different target directories in your monorepo, edit response-template.md to match your team's PR template standards, or adjust agent-narrow-component-props.yml to change trigger conditions or runner configurations. The skill's architecture in humanlayer/skills expects these files to be adapted to individual project needs while maintaining the core file naming conventions referenced in SKILL.md.

Have a question about this repo?

These articles cover the highlights, but your codebase questions are specific. Give your agent direct access to the source. Share this with your agent to get started:

Share the following with your agent to get started:
curl -s "https://instagit.com/install.md"

Works with
Claude Codex Cursor VS Code OpenClaw Any MCP Client

Maintain an open-source project? Get it listed too →