Development Story Report Skill

A cross-platform skill package for generating brief, top-down wrap-up reports after a completed story, session, or workstream. The skill applies Pyramid Principle structure to help busy, non-technical readers understand what changed, why it matters, what remains open, and what should happen next.

This repository ships ready-to-use wrappers for Claude Code and Codex, plus the shared source material and local tooling used to generate and validate them.

---

Report for This Skill

Check out our development story as a boss -:) It is automatically generated using this very skill.

---

What's Included

• Claude Code skill wrapper in claude/development-story-report/
• Codex skill wrapper in .agents/skills/development-story-report/
• Shared source-of-truth skill content in skill-src/
• Builder and validation tooling in src/development_story_report_tooling/ and scripts/build_skill_wrappers.py
• Evaluation prompts in evals/development-story-report-evals.json
• Development history and planning docs in dev-docs/

---

Why Use This?

Executive-grade wrap-ups

• Lead with the answer instead of a chronology dump
• Group findings by outcomes, decisions, and implications
• Always surface material open issues
• End with one concrete next step

Better signal for non-technical readers

• Default audience is a non-technical decision-maker
• Keeps jargon low and evidence cues light
• Stays concise enough to scan quickly

Reusable across tools

• Claude Code wrapper for custom Claude skills
• Codex wrapper for .agents/skills
• Shared authoring source so both stay aligned

---

Getting Started

Clone the repository:

git clone https://github.com/fengwang/development-story-report-skill.git
cd development-story-report-skill

Install for Claude Code

Global install:

mkdir -p ~/.claude/skills
cp -r claude/development-story-report ~/.claude/skills/

Project-level install:

mkdir -p .claude/skills
cp -r claude/development-story-report .claude/skills/

Install for Codex

Global install:

mkdir -p ~/.agents/skills
cp -r .agents/skills/development-story-report ~/.agents/skills/

Project-level install:

mkdir -p .agents/skills
cp -r .agents/skills/development-story-report .agents/skills/

Codex currently scans .agents/skills from the current working directory up to the repository root, and also supports user-level skills in ~/.agents/skills. Source: OpenAI Codex Skills.

---

Example Prompts

Once installed, prompts like these should trigger the skill naturally:

• Wrap this up for leadership.
• Summarize what we found and what remains unresolved.
• Give me a short handoff note from this work.
• What changed here, in plain English?
• Create a development story report from this session.

The default output is:

• Inline Markdown
• About 300-500 words
• Written for a non-technical decision-maker
• Structured as bottom line, key outcomes, open issues, and next step

---

Repository Layout

claude/
  development-story-report/         Generated Claude Code wrapper

.agents/
  skills/
    development-story-report/       Generated Codex wrapper

skill-src/
  shared/                           Shared content fragments
  wrappers/                         Wrapper-specific metadata and intros

src/development_story_report_tooling/
  builder.py                        Wrapper generation
  validation.py                     Smoke validation helpers

scripts/
  build_skill_wrappers.py           Build and validate both wrappers

evals/
  development-story-report-evals.json

dev-docs/
  design/
  plan/
  tasks/
  progress/
  findings/

---

Development Docs

The repo’s implementation history and working documents live under dev-docs/:

• dev-docs/design/
• dev-docs/plan/
• dev-docs/tasks/
• dev-docs/progress/
• dev-docs/findings/

The repo also includes a self-generated wrap-up report in docs/.

---

Contributing

If you update the shared reporting method:

1. Edit the shared source in skill-src/
2. Run the unit tests
3. Rebuild the generated wrappers
4. Review the emitted SKILL.md files in claude/ and .agents/

This repo intentionally favors a clean generated-source model over manual duplication or compatibility layers.

---

License

This repository is MIT-licensed. See LICENSE.