[high] fix: [btc_steroids] filter output transactions on their own value field

[elhoim]

BLUF — btc_steroids reads prev_out on output entries, so its zero-value filter never excludes anything.

• Problem — The output loop in btc_steroids.py reads tx["prev_out"]["value"], but prev_out exists only on input entries, so the lookup always raises KeyError and the zero-value filter never excludes anything.
• Fix — Reads tx["value"], the field actually present on out entries and already used to compute the displayed value.
• Effect — Analysts enriching a bitcoin address now get real payments without zero-value OP_RETURN and dust outputs cluttering the enrichment result.

In btc_steroids.py, the loop over output transactions reads tx["prev_out"]["value"]:

for tx in transactions["out"]:
    ...
    try:
        prev_out = tx["prev_out"]["value"]
    except KeyError:
        prev_out = None
    if prev_out != 0 and addr_out == btc:

Entries in transactions["out"] never have a prev_out key — that field belongs to the input entries handled in the loop above (transactions["inputs"]). So the lookup always raises KeyError, prev_out is always set to None, and prev_out != 0 is always True. The zero-value filter therefore never filters anything.

Impact: an analyst enriching a bitcoin address/transaction via the btc_steroids expansion module gets zero-value outputs (e.g. OP_RETURN or dust outputs with value == 0) listed alongside real payments, cluttering the enrichment result with entries the filter was meant to exclude.

Fix: read tx["value"] directly, which is the field actually present on out entries (and is already the field used a few lines below to compute and print the displayed value), so the zero-value filter now works as intended.

No behaviour change beyond restoring the intended filtering (fewer, correctly-filtered lines in output; no API or module contract change).

Verification

• flake8 on the changed file: clean, no output.
• Module test suite: 161 passed, 4 skipped, 5 subtests passed in 32.77s.

Found during a review of the repository; other findings are being submitted as separate PRs.

🤖 Generated with Claude Code