Agent StoreInformation TechnologyDeveloper Experience
Live

Developer Portal Content Curation Agent

Information TechnologyDeveloper Experience

Keeps internal developer portal documentation accurate and current by detecting stale content, broken links, and undocumented services.

4
Process steps
6
Integrations
3
Data inputs

Internal developer portals meant to serve as the single source of truth for service catalogs, API documentation, and onboarding guides decay quickly because documentation ownership is diffuse and updating docs rarely feels as urgent as shipping features

Engineers waste significant time following outdated setup instructions, calling deprecated API endpoints still listed as current, or searching for services that were never documented at all

Portal content that references retired systems or old team ownership actively misleads new engineers rather than helping them, undermining trust in the portal to the point where teams stop consulting it altogether

Without active curation, the gap between what the portal says and what is actually true keeps widening, and no one owns the job of closing it

The agent continuously compares developer portal content against the live state of the service catalog, API gateway, and ownership records to detect documentation that has drifted out of sync, including stale setup instructions, broken internal links, and deprecated endpoints still marked current. It identifies services registered in the catalog with missing or thin documentation and drafts starter content pulled from the service's code repository and API schema for a human to review and publish. The agent also tracks portal content freshness by team and surfaces a prioritized curation backlog so documentation debt gets addressed systematically rather than ad hoc.

1

Content and Source-of-Truth Comparison

  • Scan portal content against the live service catalog and API gateway
  • Detect references to deprecated or retired services
  • Check internal and external links for validity
  • Flag documentation pages with no recent update activity
Outcome: A complete list of portal content that has drifted out of sync with reality.
2

Documentation Gap Identification

  • Identify catalog services with missing or thin documentation
  • Assess documentation completeness against a defined standard
  • Prioritize gaps by service usage and criticality
  • Identify the correct owning team for each gap
Outcome: A prioritized list of undocumented or under-documented services needing attention.
3

Draft Content Generation

  • Pull service metadata, code comments, and API schemas as source material
  • Draft starter documentation for missing or thin pages
  • Flag outdated sections in existing pages with suggested corrections
  • Route drafts to the owning team for review and publish
Outcome: Ready-to-review draft content that dramatically reduces the effort to fill gaps.
4

Freshness Tracking and Reporting

  • Track content freshness scores by team and portal section
  • Maintain a prioritized curation backlog
  • Report documentation health trends to engineering leadership
  • Recognize teams maintaining consistently current documentation
Outcome: Sustained portal accuracy through visible, ongoing accountability.
Backstage
Confluence
GitHub
Swagger/OpenAPI registry
Slack
Notion