Agent discovery

Phase 5 taught coding models the public API. This page is the retrieval surface: the files directories and crawlers ingest. Context7 and skill indexes still need a maintainer submission; the MCP Registry is published from the tagged release workflow after PyPI upload.

Hosted machine indexes

File	URL
Short index	https://datenoio.github.io/iterabledata/llms.txt
Copy-paste recipes	https://datenoio.github.io/iterabledata/llms-full.txt
Well-known copy	https://datenoio.github.io/iterabledata/.well-known/llms.txt

Regenerate committed copies with python dev/scripts/generate_llms_txt.py.

MCP Registry

Root server.json describes the stdio server:

pip install iterabledata[mcp]
iterable-mcp

The release workflow publishes this file to the MCP Registry after each tagged PyPI upload (mcp-publisher login github-oidc), retrying until PyPI indexes the new version. Schema: 2025-12-11. Server name: io.github.datenoio/iterabledata (listed as of 1.0.23). PyPI ownership is the README marker <!-- mcp-name: io.github.datenoio/iterabledata -->. Setup docs: MCP.

Portable skill

Copy skills/iterabledata/SKILL.md into another repository (Cursor/Claude Code skills directory). Optionally list it on skills.sh or similar skill indexes.

Context7 / llms.txt directories

Point Context7, llmstxt.org, and similar indexes at the hosted llms.txt URL above and the GitHub repo datenoio/iterabledata. The site robots.txt allows those paths.

Canonical imports (do not drift)

from iterable import open_iterable
from iterable.convert import convert

PyPI package: iterabledata. Import package: iterable.