agentleFS
brpazbrpaz/agent-skills/skills/bspec-story-sync/SKILL.md

bspec-story-sync

Skill1 starsChanged 6 months agoMITView on GitHub

Synchronization skill for Bruno's personal spec workflow. Decomposes the business PRD into small vertical stories and reconciles them into GitHub issues with stable labels, bodies, and safe update rules.

We split this skill into the rules it gives an agent

31 rules. About 0 are common, 21 are distinctive, and 10 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈21

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 10

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill only for converting the business PRD into GitHub story issues.

    line 8
  2. Distinctive:

    Do not create local story tracking files.

    line 16
  3. Very specific:

    write issue links back into the relevant PRD file inside prd_document

    line 26
  4. Distinctive:

    prefer vertical slices over horizontal tasks

    line 32
  5. Very specific:

    keep product scope anchored in prd_document

    line 35
  6. Very specific:

    use approved architecture_document guidance to preserve technical constraints, not to redefine product scope

    line 36
  7. Distinctive:

    use the shared workflow contract for story IDs, managed labels, status labels, and structured markers

    line 37
  8. Very specific:

    use gh issue for issue creation and editing

    line 71
  9. Very specific:

    use gh label when workflow labels from the shared workflow contract do not exist yet

    line 72
  10. Very specific:

    prefer one area/<name> label when the PRD clearly implies a primary area

    line 74
  11. Distinctive:

    Create labels from the shared workflow contract if they do not exist yet:

    line 79
  12. Very specific:

    Use skills/bspec-story-sync/issue-template.md as the starting body template and rewrite the placeholders for the actual story.

    line 87
  13. Distinctive:

    When the story already exists and is safe to update:

    line 101
  14. Very specific:

    start from skills/bspec-story-sync/issue-template.md and replace the example content with story-specific content before publishing

    line 116
  15. Distinctive:

    keep acceptance criteria executable and product-facing

    line 117
  16. Distinctive:

    if a related ADR exists, link it without copying its contents

    line 121
  17. Distinctive:

    prefer concise issue bodies over long design documents

    line 122
  18. Distinctive:

    ensure there is exactly one workflow status label on the issue

    line 123
  19. Distinctive:

    create a new issue when the story ID does not exist

    line 127
  20. Distinctive:

    update an issue only when it is open and not actively being implemented or reviewed

    line 128
  21. Distinctive:

    never silently overwrite in-progress, in-review, or closed issues

    line 129
  22. Very specific:

    after creating or matching an issue, write the issue reference back into the owning PRD file inside the logical prd_document

    line 134
  23. Distinctive:

    keep product intent and issue execution state separate: the PRD stores references, GitHub stores the executable story state

    line 135
  24. Distinctive:

    place links near the relevant feature or story section inside the PRD file that owns it

    line 136
  25. Very specific:

    prefer a compact subsection such as #### GitHub Issues

    line 137
  26. Distinctive:

    use markdown links with both story ID and issue number when available

    line 138
  27. Distinctive:

    if the story is project-level, place the linkback in the PRD index

    line 139
  28. Distinctive:

    prefer issue state plus labels over local workflow files

    line 152
  29. Distinctive:

    keep issue bodies concise and executable

    line 153
  30. Distinctive:

    use the shared workflow contract as the source of truth for labels, status names, story IDs, and markers

    line 154
  31. Distinctive:

    keep the PRD logically unified even when feature sections live in separate files

    line 156

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Story Sync

Use this skill only for converting the business PRD into GitHub story issues.

Source Of Truth

  • prd_document = business and product truth, stored in one file or in an index plus section files
  • architecture_document = optional approved technical architecture guidance for issue references and constraints
  • GitHub issues = story truth

Do not create local story tracking files.

Responsibilities

  • extract story-ready work from the PRD
  • split large features into vertical slices
  • format concise executable issue bodies
  • create missing issues
  • update safe-to-update issues
  • bootstrap labels when needed
  • write issue links back into the relevant PRD file inside prd_document
  • preserve useful architecture references on issues
  • report drift on active or closed issues

Story Rules

  • prefer vertical slices over horizontal tasks
  • each story should have one meaningful outcome
  • each story should be small enough for one focused implementation cycle
  • keep product scope anchored in prd_document
  • use approved architecture_document guidance to preserve technical constraints, not to redefine product scope
  • use the shared workflow contract for story IDs, managed labels, status labels, and structured markers

Issue Template

Story-ID: US-001

## Story
As a <user>, I can <action> so that <outcome>.

## Acceptance Criteria
- ...

## Non-Goals
- ...

## Notes
- concise product-facing notes

## Technical Notes
- relevant architecture constraints or references

## Testing Notes
- relevant test cases or strategies

## PRD Trace
- prd_document: <file path> :: <section reference>

## Architecture Trace
- architecture_document: <section reference>

GitHub Execution Rules

  • use gh issue for issue creation and editing
  • use gh label when workflow labels from the shared workflow contract do not exist yet
  • apply the managed label, story label, and exactly one status label from the shared workflow contract
  • prefer one area/<name> label when the PRD clearly implies a primary area
  • preserve issue discussion history when updating existing issues

Label Bootstrap Example

Create labels from the shared workflow contract if they do not exist yet:

bash skills/bspec-story-sync/ensure-bspec-labels.sh

Issue Creation Example

Use skills/bspec-story-sync/issue-template.md as the starting body template and rewrite the placeholders for the actual story.

gh issue create \
  --title "[US-001] User can create a project" \
  --label "bspec" \
  --label "story" \
  --label "status/ready" \
  --label "area/project-setup" \
  --body-file skills/bspec-story-sync/issue-template.md

Issue Update Example

When the story already exists and is safe to update:

gh issue edit 123 \
  --title "[US-001] User can create a project" \
  --add-label "bspec" \
  --add-label "story" \
  --add-label "status/ready" \
  --add-label "area/project-setup" \
  --body-file /tmp/us-001-issue.md

Issue Body Rules

  • keep Story-ID stable
  • start from skills/bspec-story-sync/issue-template.md and replace the example content with story-specific content before publishing
  • keep acceptance criteria executable and product-facing
  • include a PRD Trace section that points back to the canonical PRD file and section
  • include an Architecture Trace section when relevant approved architecture guidance exists
  • preserve lightweight Architecture and ADRs reference sections when updating existing issues
  • if a related ADR exists, link it without copying its contents
  • prefer concise issue bodies over long design documents
  • ensure there is exactly one workflow status label on the issue

Reconcile Rules

  • create a new issue when the story ID does not exist
  • update an issue only when it is open and not actively being implemented or reviewed
  • never silently overwrite in-progress, in-review, or closed issues
  • report drift clearly instead of forcing synchronization

PRD Linkback Rules

  • after creating or matching an issue, write the issue reference back into the owning PRD file inside the logical prd_document
  • keep product intent and issue execution state separate: the PRD stores references, GitHub stores the executable story state
  • place links near the relevant feature or story section inside the PRD file that owns it
  • prefer a compact subsection such as #### GitHub Issues
  • use markdown links with both story ID and issue number when available
  • if the story is project-level, place the linkback in the PRD index

Example:

#### GitHub Issues
- [US-001 / #123](https://github.com/example/repo/issues/123) User can create a project

Rules

  • preserve GitHub discussion history
  • preserve existing lightweight Architecture and ADRs references on safe-to-update issues
  • prefer issue state plus labels over local workflow files
  • keep issue bodies concise and executable
  • use the shared workflow contract as the source of truth for labels, status names, story IDs, and markers
  • replace old workflow status labels when changing issue state; do not accumulate multiple status labels
  • keep the PRD logically unified even when feature sections live in separate files

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.