Skip to content

docs: competitive upgrade plan vs OSSInsight + TrendShift (2026-06-14) - #3171

Open
0motionguy wants to merge 1 commit into
mainfrom
docs/competitive-upgrade-plan-2026-06-14
Open

0motionguy wants to merge 1 commit into
mainfrom
docs/competitive-upgrade-plan-2026-06-14

Conversation

@0motionguy

Copy link
Copy Markdown
Owner

Summary

  • Plan markdown driven by a 3-way side-by-side scrape of trendingrepo.com, ossinsight.io, trendshift.io through the new Toolbox engine bench (companion PR: 0motionguy/toolbox#239).
  • 15 punch-list items across 4 sections: Freshness wins (F1–F5), Coverage wins (C1–C4), UX wins (U1–U4), LLM consensus restoration (L1–L2). Each has a Why / How (file pointer) / Verify.
  • Phase 0 (blocker): the F1–F3 cron tightening targets the velocity-engine fetchers which only exist on the unmerged hardening/2026-05-31-wave branch. The originally-planned companion code PR was deferred until branch reconciliation.
  • All competitor data was gathered through OUR engines (Firecrawl + Toolbox web.scrape.v1 + Context API + worker-cloak) — never via WebSearch. Drink-our-own-champagne move that also serves as live validation of the 2026-06-14 scrape engine upgrades.

What the scrape showed

TrendingRepo OSSInsight TrendShift
Per-repo depth wins (10-component momentum + breakout + mentions) dense charts shallow
Freshness signal in DOM weak weak strong ("Today (UTC)")
Collection depth medium wins (102 curated) 40 topics
API exposure wins (REST + MCP + Portal v0.1) public API none
LLM narratives broken (Kimi 403) none none
Discovery filters visible on home weak medium strong

Tangible gaps to close: (1) explicit freshness markers on the homepage, (2) language-filter visibility, (3) the dead LLM consensus.

Test plan

  • Read the markdown end to end — confirm the four-phase order makes sense
  • Decide on Phase 0: merge hardening/2026-05-31-wave to main OR target F1–F3 directly to the hardening branch
  • Pick the top 5 items to ship in the next TR session

🤖 Generated with Claude Code

Punch list derived from a side-by-side Toolbox engine scrape of
trendingrepo.com, ossinsight.io, and trendshift.io. The scrape itself
ran through the new in-repo bench harness at toolbox PR #239 (Wave 1
engine performance + Wave 2 competitive feature extraction). All
competitor data came through our engines — Firecrawl, Toolbox
web.scrape.v1, Context API, cloak — not WebSearch.

Four sections, each item scoped <1 day with Why / How (file pointer)
/ Verify lines:

  1. Freshness wins (F1-F5)         — cron tightening + homepage stamp
  2. Coverage wins (C1-C4)          — Reddit, TrendShift, video, hourly
  3. UX wins (U1-U4)                — language chips, time windows, etc.
  4. LLM consensus restoration (L1-L2) — NanoGPT swap (Kimi dead)

Phase 0 (blocker) — F1-F3 cron edits target the velocity-engine
fetchers (velocity-refresh, velocity-backfill, star-activity-deltas)
which live only on the unmerged hardening/2026-05-31-wave branch. The
companion freshness PR was deferred until branch reconciliation
lands. Everything else in the plan is actionable today.

Sister deliverable: 0motionguy/toolbox#239 (scrape-bench Wave 1 + Wave 2).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant