Found while merging #89 into #109.
extension/shared/consensus-client.js makes up a consensus figure when Consensus doesn't return one:
- In
verifyFindingsWithConsensus, a search that returns papers but no consensus_meter gets meter = 85, and one with no papers gets 50: data.consensus_meter?.yes_pct ?? (papers.length > 0 ? 85 : 50).
- In
formatConsensus, a cached record without a numeric meter gets 88, and totalStudies defaults to 1.
The side panel then renders this as "✓ 85% Consensus (5 papers)". It reads as a measured agreement level, but it is a constant. idstack's premise is honest evidence labels, so this should probably render as "✓ Consensus match" (or show no badge) when the API gives no meter.
The same question applies to bin/idstack-consensus. Check whether cmd_verify and cmd_query also substitute defaults.
🤖 Generated with Claude Code
Found while merging #89 into #109.
extension/shared/consensus-client.jsmakes up a consensus figure when Consensus doesn't return one:verifyFindingsWithConsensus, a search that returns papers but noconsensus_metergetsmeter = 85, and one with no papers gets 50:data.consensus_meter?.yes_pct ?? (papers.length > 0 ? 85 : 50).formatConsensus, a cached record without a numericmetergets 88, andtotalStudiesdefaults to 1.The side panel then renders this as "✓ 85% Consensus (5 papers)". It reads as a measured agreement level, but it is a constant. idstack's premise is honest evidence labels, so this should probably render as "✓ Consensus match" (or show no badge) when the API gives no meter.
The same question applies to
bin/idstack-consensus. Check whethercmd_verifyandcmd_queryalso substitute defaults.🤖 Generated with Claude Code