Release Notes Expert
Overview
Transform raw technical changes -- tickets, changelogs, git logs, PRDs -- into clear, user-benefit-oriented release notes. This skill ensures every release communicates value to the right audience in the right tone.
When to Use
- Product Releases -- Announcing new versions to customers, partners, or internal stakeholders.
- Sprint Demos -- Summarizing what shipped for sprint review audiences.
- Changelog Maintenance -- Keeping a running log of changes across releases.
- Customer Communication -- Preparing release announcements for email, in-app, or documentation.
Methodology
Step 1: Gather Raw Input
Collect all changes from the release cycle:
- Jira/Linear tickets -- Completed stories, bugs, and tasks
- Git log -- Merge commits since last release tag
- PRD references -- Feature specs that shipped
- Hotfix records -- Emergency fixes deployed between releases
Step 2: Classify Each Change
Assign every change to exactly one category:
| Category | Definition | Example | |----------|-----------|---------| | New Features | Net-new capabilities that did not exist before | New export-to-PDF option | | Improvements | Enhancements to existing functionality | Faster dashboard loading | | Bug Fixes | Corrections to broken or incorrect behavior | Fixed login redirect loop | | Breaking Changes | Changes that require user action to adapt | API v2 replaces v1 endpoints | | Deprecations | Features scheduled for future removal | Legacy CSV import will be removed in v4.0 |
Classification rules:
- If a change adds something entirely new, it is a New Feature.
- If it makes something existing better (faster, easier, more reliable), it is an Improvement.
- If it fixes something that was wrong, it is a Bug Fix.
- If users must change their behavior, configuration, or integration, it is a Breaking Change.
- If a feature still works but will be removed later, it is a Deprecation.
Step 3: Rewrite for User Benefit
The most critical step. Every entry must lead with the benefit to the user, not the technical change.
Rewriting principles:
- Lead with the outcome. What can the user do now that they could not before, or what is better for them?
- Use plain language. Avoid internal jargon, code references, or implementation details.
- Keep it to 1-3 sentences. One sentence for minor items, up to three for significant features.
- Include context when needed. If users need to take action, tell them exactly what to do.
Before and after examples:
| Technical (Bad) | User-Benefit (Good) | |----------------|-------------------| | Implemented Redis caching layer for dashboard queries | Dashboards now load up to 3x faster | | Refactored authentication module to use OAuth 2.0 PKCE flow | Sign-in is now more secure and works reliably across all browsers | | Fixed null pointer exception in report export handler | Report exports no longer fail when date ranges include empty days | | Migrated user preferences API from v1 to v2 schema | Action required: Update your API calls to use the new /v2/preferences endpoint by April 30. See migration guide. | | Added feature flag for beta dashboard | You can now opt into the redesigned dashboard from Settings > Beta Features |
Red flags that an entry needs rewriting:
- Mentions a class name, function, or library
- Starts with "Refactored," "Migrated," or "Updated" without stating impact
- Uses acronyms the target audience would not know
- Describes what the team did instead of what the user gains
Step 4: Adjust Tone for Audience
| Audience | Tone | Style Notes | |----------|------|-------------| | B2B / Enterprise | Professional, precise | Emphasize reliability, security, compliance. Avoid casual language. | | Consumer | Friendly, conversational | Use "you" and "your." Celebrate new features. Keep it light. | | Developer / API | Technical, direct | Include endpoint names, SDK versions, code snippets. Be specific. | | Internal | Detailed, context-rich | Include ticket IDs, team names, technical details as needed. |
Step 5: Assemble the Release Notes
Use the output template below. Include only categories that have entries -- do not show empty sections.
Output Template
# [Product Name] v[X.Y.Z] Release Notes
**Release Date:** [YYYY-MM-DD]
---
## New Features
- **[Feature Name]** -- [1-3 sentence description of user benefit]. ([TICKET-ID])
## Improvements
- **[Improvement Name]** -- [1-2 sentence description of what is better]. ([TICKET-ID])
## Bug Fixes
- **[Bug Fix Name]** -- [1 sentence describing what was broken and that it is now fixed]. ([TICKET-ID])
## Breaking Changes
> **Action Required:** The following changes require updates on your end.
- **[Change Name]** -- [Description of what changed and exactly what the user must do]. ([TICKET-ID])
## Deprecations
> **Planned Removal:** The following features will be removed in a future release.
- **[Feature Name]** -- [What is being deprecated and when it will be removed. Recommend alternative if available]. ([TICKET-ID])
---
**Full changelog:** [link]
**Questions?** [support link or contact]
Python Tool
Use scripts/release_notes_generator.py to generate formatted release notes from structured input.
# Generate from JSON input
python scripts/release_notes_generator.py --input changes.json --product-name "Acme App" --version "2.5.0"
# Run with demo data
python scripts/release_notes_generator.py --demo --product-name "Acme App" --version "1.0.0"
# Output as JSON instead of markdown
python scripts/release_notes_generator.py --input changes.json --format json --product-name "Acme App" --version "2.5.0"
See scripts/release_notes_generator.py --help for full usage.
Integration with Other Skills
- Use
summarize-meeting/to capture release planning discussions. - Use
job-stories/orwwas/to trace features back to their original motivation. - Pair with
../senior-pm/for stakeholder communication planning around major releases.
Troubleshooting
| Symptom | Likely Cause | Resolution | |---------|-------------|------------| | Tool flags too many entries as "technical language" | TECHNICAL_PATTERNS regex is broad, catching common words like "update" or "add" | Review the flagged entries; the tool provides suggestions, not mandates -- ignore false positives for user-facing descriptions | | All entries classified as same type | Input JSON uses wrong type values or inconsistent casing | Use exact lowercase types: feature, improvement, bugfix, breaking, deprecation | | Empty sections appear in output | No entries of that type exist but template still renders the section | The tool only renders sections with entries; empty sections indicate a data issue in input | | Breaking changes not highlighted prominently | Entries marked as improvement instead of breaking | Review classification rules: if users must change behavior, config, or integration, it is breaking, not improvement | | Release notes sound like commit messages | Descriptions written from developer perspective, not user perspective | Apply the rewriting principles: lead with outcome, use plain language, 1-3 sentences per entry | | --demo flag requires --product-name and --version | These are required arguments regardless of input source | Always provide both: --demo --product-name "App" --version "1.0.0" | | JSON output missing rewriting suggestions | No entries triggered technical language detection patterns | This is expected behavior; suggestions only appear when TECHNICAL_PATTERNS match entry descriptions |
Success Criteria
- Every release has structured notes published within 24 hours of deployment
- All entries lead with user benefit, not technical implementation details
- Breaking changes include explicit migration instructions with deadl