leo: ingest skill #2988

Closed
m3taversal wants to merge 11 commits from leo/ingest-skill into main

11 commits

Author SHA1 Message Date
a25ef69ffa leo: fix pending file cleanup — failed extractions no longer block retries
Failed/timed-out extractions and no-change extractions now remove the
source from extract-pending.txt so the next cron cycle can retry.

Caught by: Ganymede <00F28B10-062E-4863-9DD2-A5E9407B33FA>
Pentagon-Agent: Leo <14FF9C29-CABF-40C8-8808-B0B495D03FF8>
2026-03-10 10:52:11 +00:00
9987925ced leo: fix critical issues from Ganymede's code review
Fixes:
- CRITICAL: Add --permission-mode bypassPermissions (Claude would hang on headless)
- CRITICAL: Track pending extractions in extract-pending.txt to prevent re-processing loop
- WARNING: Use jq for PR JSON construction (prevents injection from filenames)
- WARNING: Add duplicate PR guard (check before creating)
- WARNING: Stage only files in inbox/archive/ and domains/ (prevents accidental inclusions)
- WARNING: Use git credential helper instead of token in URL (keeps tokens out of logs)
- MINOR: Fix flagged_for example in ingest.md
- MINOR: Add rejected-extraction guidance to Track B

Reviewed by: Ganymede <00F28B10-062E-4863-9DD2-A5E9407B33FA>
Pentagon-Agent: Leo <14FF9C29-CABF-40C8-8808-B0B495D03FF8>
2026-03-10 10:48:55 +00:00
751a651ce8 Auto: ops/extract-cron.sh | 1 file changed, 96 insertions(+), 27 deletions(-) 2026-03-10 10:48:26 +00:00
0c83c78132 leo: document concurrency model and domain routing in extract-cron
Pentagon-Agent: Leo <14FF9C29-CABF-40C8-8808-B0B495D03FF8>
2026-03-10 10:39:37 +00:00
Leo
6d68c2c417 Merge branch 'main' into leo/ingest-skill 2026-03-10 10:34:14 +00:00
71ec1af778 Auto: skills/ingest.md | 1 file changed, 59 insertions(+), 24 deletions(-) 2026-03-10 10:32:53 +00:00
e35e894240 leo: split ingestion — agents research + archive, VPS extracts headlessly
- What: Rewrote skills/ingest.md to be research-only (find sources, archive with notes)
- Added ops/extract-cron.sh — VPS cron job that picks up unprocessed sources, runs Claude headless to extract claims, opens PRs
- Why: Separates high-judgment work (research) from mechanical work (extraction). Agents spend session time finding sources, not grinding through extraction. Archive everything regardless of whether claims come out.
- Architecture: Agents archive → VPS extracts → VPS eval reviews → auto-merge

Pentagon-Agent: Leo <14FF9C29-CABF-40C8-8808-B0B495D03FF8>
2026-03-10 10:31:49 +00:00
e991e55ab7 Auto: ops/extract-cron.sh | 1 file changed, 167 insertions(+) 2026-03-10 10:31:39 +00:00
fb18064f97 Auto: skills/ingest.md | 1 file changed, 83 insertions(+), 109 deletions(-) 2026-03-10 10:31:02 +00:00
63089abe63 Auto: skills/ingest.md | 1 file changed, 192 insertions(+) 2026-03-10 10:23:34 +00:00
c9e2970cfb Auto: 3 files | 3 files changed, 677 insertions(+), 81 deletions(-) 2026-03-09 22:26:36 +00:00