Skip to content

Garbage owner URLs write permanent owners rows on 404 #260

New issue
New issue
Closed
#262
Closed
Garbage owner URLs write permanent owners rows on 404#260
#262
Labels
harlan-agent-ready-to-implementIssue triage found bounded work ready for an implementation Agent.routine:daily-checkin

Description

@harlan-github-agent
harlan-github-agent
bot
opened on Sep 20, 2026

The org profile route inserts a permanent owners row with sync_status='404' for every raw owner path GitHub answers 404 for, so crawler-invented garbage URLs grew the owners table 2,941 junk rows in 24h (6,775 total since ~2026-09-17, 9,868 owners with no repos row) while each miss also pays one unauthenticated GitHub API call.

Target: layers/registry/server/api/orgs/[owner].get.ts

Verify with: pnpm dlx wrangler d1 execute skilld-db --remote --json --command "SELECT COUNT(*) FROM owners WHERE sync_status='404' AND last_synced_at >= strftime('%s','now')-86400" returns near zero for the 24h after the fixed route deploys, and curl -s -o /dev/null -w '%{http_code}' 'https://skilld.dev/api/orgs/not-a-real-owner-xyz' stays 404 with no new owners row.

Estimated to change 2 files.

The daily-checkin routine opened this issue automatically on its 2026-09-20T21:40:00.000Z run. It is not Harlan's own report. Close it to reject the proposal, and the reason you give stops it being offered again.

Activity

  1. harlan-github-agent
    added
    routine:daily-checkin
    harlan-agent-runningAn Agent holds a Task on this issue or pull request right now.
    on Sep 20, 2026

  2. harlan-github-agent commented on Sep 20, 2026

    @harlan-github-agent
    harlan-github-agentBot
    on Sep 20, 2026 – with Harlan GitHub Agent
    ContributorAuthor

    🤖 ISSUE TRIAGE

    Harlan Agent Kit posted this automated triage. It is not Harlan's personal assessment or commitment. AI open source policy.

    • Route: Ready to implement
    • Difficulty: 2/5
    • Impact: 4/5
    • Reproduction: Yes
    • Codebase review: Not needed
    • Summary: Confirmed in layers/registry/server/api/orgs/[owner].get.ts: fetchAndStoreOwner upserts a permanent owners row with sync_status='404' whenever GitHub answers 404, and loadOrgProfile calls loadOwner in parallel before discovering the registr
    • Next action: In loadOrgProfile, sequence querySkills before loadOwner (drop the Promise.all), throw the existing 404 for an empty result before calling loadOwner, and add a regression test asserting a garbage owner triggers no GitHub fetch and no owners row; optionally add a storage-based negative cache instead of the '404' owners row for owners that lose their GitHub account.
  3. harlan-github-agent
    added
    harlan-agent-ready-to-implementIssue triage found bounded work ready for an implementation Agent.
    harlan-agent-runningAn Agent holds a Task on this issue or pull request right now.
    and removed
    harlan-agent-runningAn Agent holds a Task on this issue or pull request right now.
    on Sep 20, 2026
  4. harlan-zw
    closed this as completedin #262on Sep 21, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    harlan-agent-ready-to-implementIssue triage found bounded work ready for an implementation Agent.routine:daily-checkin

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions