API sync authoring skill
Install chkit-ingestion to give your coding agent source-authoring instructions and links to the relevant guides.
Install
Section titled “Install”From the application project:
npx skills add obsessiondb/chkit --skill chkit-ingestionThe TypeScript CLI also provides a pass-through:
chkit skills add obsessiondb/chkit --skill chkit-ingestionChoose the agent in the installer. Install @chkit/plugin-ingest and configure credentials through the quickstart before running a sync. The separate chkit skill covers schema and migration workflows.
To test unpublished changes to the skill, install it from a local checkout of the chkit repository:
npx skills add ./skills/chkit-ingestionUse it
Section titled “Use it”For an agent supporting explicit skill invocation:
Use $chkit-ingestion to add a source for our helpdesk tickets.Inspect the existing schema and provider documentation first.Explain the choices for stored shape, transformations, incremental state,and loading, then implement the reader and tests.Include provider documentation, example responses, freshness needs, and any restrictions on storing raw fields. The agent should derive decisions from those facts and the existing project.
Skill scope
Section titled “Skill scope”The skill covers bounded readers, provider-specific sync strategies, durable checkpoint ordering, write identity, and scheduling. For official registry integrations, it requires a source contract and recovery design, raw payload preservation, late-change reconciliation, and portable interruption/replay tests. Sync tokens, timestamp windows, and checkpointed full scans follow the provider’s actual capabilities. Use the linked docs for complete examples and tradeoffs.
Related pages
Section titled “Related pages”- Destinations and transformations: choose storage and mapping for the source.
- Quickstart: a working first source.
- Publish a registry: metadata, portable tests, and immutable releases.
- For AI agents: general chkit setup and schema workflows.