[harlan-zw]

🔗 Linked issue

Related to #179.

❓ Type of change

• 📖 Documentation
• 🐞 Bug fix
• 👌 Enhancement
• ✨ New feature
• 🧹 Chore
• ⚠️ Breaking change

📚 Description

X discovery cannot use a remainder below its minimum request size, but the check-in missed that exhaustion.
Read the request minimum from the same source as the ceiling and flag unusable remainders.
Record distinct X Posts within the observation day and preserve unknown flags when the source is incomplete.
Report instructions now describe projected spend as a Post-only estimate with actual billing unavailable.

🤖 AI disclosure: Harlan Agent Kit modified this description. My AI open-source policy.

[harlan-zw]

Harlan Agent Kit wrote this verification note.

Read-only production queries on 2026-10-05 showed 1,992 returned discovery reads today.
Only eight reads remained, below X search's ten-result minimum.
The old check-in flag reported false for this input.

D1 snapshots filtered to X recorded 498 distinct Posts on October 4 and 499 on October 5, a partial UTC day.
This corrects the earlier counts, which included Bluesky in the shared snapshot table.
These combine discovery and refresh. They do not identify actual billed usage or author-expansion charges.
X pricing describes soft deduplication within each UTC day.
Treat the existing dollar projection as a Post-only estimate.

#409 fixes repeated first-page discovery reads independently.
Measure complete UTC days after it lands before changing the query or ceiling for #179.

[harlan-github-agent]

🤖 MERGED

Harlan Agent Kit posted this automated review. It is not Harlan's personal review or approval. AI open source policy. Last updated: 2026-10-05 07:52 UTC.

GitHub merged this pull request.

No material findings were recorded.

7cab0470-156f-41fb-91b7-ea2a442a9a36

[harlan-zw]

🤖 Harlan Agent Kit Agent assessment of head 6d91dda3c1b302fa82d5d760f6540a02c35b03e6.

The Service reported CapacityUnavailable. Its exact-head Review was stopped before an independent Agent reviewed the complete diff.

The review found no defects. All 15 focused tests passed. Confidence: 95/100, with five points deducted for focused coverage.

This is an independent Agent assessment. The Service Review for this head remains stopped.

[harlan-github-agent]

Harlan Agent Kit posted this automated review. AI open source policy.

Automated Review: READY.

Read the Review results