Delta Extractor

Who the node is. The id is the key the DOT pins.

id

delta-extractor

name

Delta Extractor

type

tool

phases

none declared

What it does, and the prose the agent is handed when the graph runs.

action30 words

Query the warehouse for every row changed inside the window and return them as a table, paging at the row cap so one runaway night cannot pull the whole history.

spec107 words

Query the warehouse for every row whose change timestamp falls inside the window you receive on window, reading in pages of 5000 rows ordered by that timestamp so paging stays stable across the run. Emit the rows on rows over a fixed set of columns, in the order the warehouse returned them, with no reshaping or filtering of your own. Stop at 50000 rows and report the timestamp you stopped at, so a runaway night comes back truncated and resumable instead of pulling the whole history. Issue reads only: you must not write, update or delete anything in the warehouse, whatever the window turns out to contain.in full above

model

whatever the graph supplies

agent

Delta extractor

skill

skills/delta-extractor.md

tools

SQL

mcp

postgres

params

max_rows: 50000, page_size: 5000

What arrives, what leaves, which nodes it expects to hear from, and what may not.

inputs

window : json

outputs

rows : table

dependencies

delta-intake

cannot

no type is refused

will_not

write anything back to the warehouse, reshape the rows it returns

The keys the static analysis reads. Nothing here instructs the agent.

risk_markers

none

notes60 words

This node only reads, but the vocabulary's sql capability does not distinguish a read from a write and the analyzer cannot tell the difference. Under ontology v0.1 that shows up as an inferred unvalidated-external-access on every sql node whose consumer is not a validation node, here it is arguably fair, since the rows really do come from outside the graph.

The card's own version, and who wrote it.

version

1.0.0

author

autogen

provenance

not stated