research
deep dossiers — receipts, verdicts, doctrine inputs
Doctrine for this doc: delve into the actual code, not the authors' words. Every pattern below is grounded in a real file path from a clo…
Three lanes (docs/RESEARCH_postcutoff_pulse.md · RESEARCH_frontier2.md · RESEARCH_cli_capabilities.md), synthesized.
*Focused research for the 2-hour sprint after the ~10× cost-accuracy fix (78ed19a: streaming
Source: "Peter Steinberger | OpenClaw Creator," Greg Kamradt's channel (82YaJw-_t10),
*2026-07-07 · Fable integration agent · 138 repos judged (60 steipete + 78 openclaw org) · winners installed with evidence*
Date: 2026-07-07 · Method: 5 parallel research lanes (sonnet subagents) over gh search/gh api
*Second surface-widening pass. 5 parallel Sonnet lanes (E–I), same rules as volume 1
Research spree lane 4. Practitioner-weighted: for each voice, the ONE most actionable idea/technique/artifact from roughly the last 60 da…
Mined against what's already captured in ~/.claude/skills/loop-engineering/SKILL.md (provenance section) and ~/Developer/claude-mission-c…
*Targeted research to de-risk building L11 hill-climb (the harness self-improvement loop). 3 Sonnet lanes: (1) self-improvement mechanics…
*Third loop-research push, on the territory the first two didn't map. 4 Sonnet lanes: (A) observability & fleet health, (B) multi-agent t…
*The wide sweep. 5 agent-reach lanes across every live backend: (1) X/Twitter topic sweep, (2) Reddit agent-builder subs, (3) YouTube con…
> A practitioner's extraction of the high-ticket, implementable ideas from Matt Pocock's (@mattpocockuk) AI-engineering era (~late 2025 →…
*What makes a repo KEEP compounding stars after the launch spike — and what KILLS traction.
Lens: the launch-day playbook for an open-source developer tool, reverse-engineered from
*2026-07-08 · OSS-traction research agent · lens: the person-behind-the-repo, the distribution flywheel, cross-pollination credibility, a…
*Lens: the single most directly-applicable question — how did tools in the EXACT same niche (Claude Code / AI-coding power-user tools) ga…
*Lens: the GitHub repo page as a conversion funnel. A stranger lands from HN / X / a "show me your setup" thread, and decides in ~5 secon…
Scheduled re-sweep (TROVE.md §6 "Re-sweep dates" → *post-cutoff re-sweep after 2026-07-08 07:00 UTC*; GAP 3 / task #47).
Sweep 2026-07-07 ~02:00–02:30 UTC (Jul 6 ~6:00–6:30pm PT) via agent-reach: X/Twitter via OpenCLI (live + top tabs,
FIXED: idle CPU 98.8% → 0%. Root cause was the App scene's .commands + MenuBarExtra
Target: oraios/serena — ~26k-star "MCP toolkit for coding: semantic retrieval and editing."
Sources, all from the AI Engineer conference channel, distilled 2026-07-17 from yt-dlp captions:
Scope: the user's linked video (hyqLNX3VExQ) plus the 10 most substantive other uploads pulled via
*Twitter/X power-user signal, gathered 2026-07-06 via agent-reach (platform Twitter, backend OpenCLI).
*Real power-user signal on Claude Code waste, cost, model-routing, and workflow bloat, gathered
Source: X Article by @Av1dlive (x.com/i/article/2074094402298130433, Jul 6 2026, 1,073♥, 1.7M views). "Written using the user's notes, ed…
Bun (JS runtime, originally Zig) was mechanically ported to Rust — 1,009,272 lines added, 6,502 commits (peak 695/hr) — in 11 days (May 3…
Source: Twitter/X via agent-reach (OpenCLI backend), 2026-07-06. 71 unique tweets across 6 queries.
Date: 2026-07-07 · Installed CLI: 2.1.202 (native binary, ~/.local/share/claude/versions/2.1.202, 243MB)
*Written 2026-07-06. Grounded in (a) Twitter power-user pain pulled live via agent-reach/OpenCLI,
Treasure-trove extraction for Claude Mission Control (a Claude-Code observability app that
diffs judged blind by GPT-5.5 as X and Y. Cost: GLM ~$0.04 vs Opus ~$2.73
Research sweep 2026-07-06 via gh api / gh search (account ss251). Companion to
Research sweep 2026-07-07 via agent-reach (X/Twitter + Reddit via OpenCLI bridge). Both channels healthy; no auth gaps.
Narrow, operational follow-up to ~/Developer/claude-mission-control/docs/RESEARCH_fable_rush.md (read that for the broad landscape). This…
> Round-1 built SEE (Attention Strip + Fleet Board), AUDIT (cache-miss / dead-skills / doctrine / mismatch), LAUNCH (Session Builder + Sk…
> Companion to RESEARCH_frontier.md (X landscape, Jul 6) and RESEARCH_fable_rush_github.md
- SuperGrok Heavy: normally $300/mo; current promo $99/mo for 3 months (67% off), started ~Jul 12 2026 → promo expiry ≈ 2026-10-12 for th…
*A different surface from the top-voices ledgers: HN threads (comment gold, not just headlines), Lobsters, technical Reddit, and the long…
Target repos (cloned to ~/Developer/reference-repos/):
Mined from the actual code (not the videos) in disler's Claude Code repos. Every pattern
Source: "Jensen Huang: Why companies need open agent systems," LangChain YouTube channel
*agent-reach (Twitter/X via OpenCLI) sweep, 2026-07-06. ~46 searches, 15 results each, 170+ unique
*Compare pass, 2026-07-09, triggered by trifola-radar run #1 surfacing that Anthropic shipped a native "which agent needs you" surface. T…