Blackeye PAC

CWP

Parse disclosure PDFs to surface self-enriching members. Feeds the Challenger Project.
HomePlanningMilestonesTasksIssuesBuild NotesLinks
Implementation write-back — discoveries that contradict the plan, and blockers that stop progress. A blocker automatically spawns a linked task in your task list.
Markdown supported — load a .md file, use the toolbar, or hit Preview to see the result.
opendiscoverywealth-summary API includes under-review years in baseline/latest (must fix before report cards)

services/conflict/report-api.js wealthSummary() reads all wealth_snapshots rows without filtering under_review (349 snapshot-years are under review). Baseline/latest/WGS on the Report Card headline endpoint can therefore be built on excluded-from-publication years — violates the project principle that under-review years are publicly excluded. Fix: exclude under_review from baseline/latest/wgs and surface an excluded-years count so cards can footnote honestly. Plan ref: docs/CWP-Pilot-and-Report-Card-Execution-Plan.md item T1.0.4/B1.

2026-08-08 · by claude-code
Edit
opendiscoveryThread 1 premise correction: pilot selection has NOT been done; verification depth is the real cost

Contrary to the assumption that "Thread 1 was just completed": no pilot list exists in the repo, MyCoFounder, or any session transcript. The 5 report cards in data/conflict-report-cards/ are the Project C technical demo (all publishable=false), not a selection. Additionally, the Thread 1 rubric's "verified corpus" filter cannot be satisfied by filtering alone — only 202 disclosures corpus-wide are verified_correct, and the stark-growth shortlist candidates each have ~0 verified filings. Thread 1 therefore includes a targeted human-verification pass over the ~15-18 shortlisted members (~1-2 hrs each via anchor-assisted review) before locking. Full plan with reproducible shortlist query: docs/CWP-Pilot-and-Report-Card-Execution-Plan.md §2 (branch docs-cwp-pilot-plan).

2026-08-08 · by claude-code
Edit