cyotee/aeon-paper-pick
lib/bankr-skills/aeon-paper-pick/SKILL.md
Surface the one AI / ML paper to read today from Hugging Face Papers, with the central claim, why it's worth an hour, where it might be wrong, and a time-budgeted read order. Filters out pure benchmark-chasing and incremental scaling reports. Use as a daily morning brief input for AI-savvy operators. Triggers: "one paper to read today", "best AI paper today", "what's the must-read paper", "HF Papers top pick".
data-ai
Updated Jul 22, 2026
$ install --global
skillsauthnpx skillsauth add cyotee/indexedex aeon-paper-pickInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: Jun 18, 2026, 3:26 AM360.4s2 files scanned
SKILL.md
- name:
- aeon-paper-pick
- description:
- |
- Triggers:
- one paper to read today", "best AI paper today", "what's the must-read paper",
aeon-paper-pick
One paper per day. Not a digest — one pick, with a short brief that lets the reader decide whether to invest the next hour.
How the pick works
- Fetch last 24h of HF Papers (
https://huggingface.co/papers). - Drop pure benchmark-chasing (new SOTA on existing leaderboard, no method change), incremental scaling reports, and position papers without empirical claims.
- Score survivors on novelty of method/framing, falsifiability of central claim, reproducibility signal (code + weights + data), and cross-discipline applicability.
- Pick the highest scoring. Tie-break: sharper falsifiable claim.
Brief format
[Paper title]
Authors: A, B, C (Lab)
arXiv: 2505.xxxxx
Central claim: [one sentence, plain English]
Why it's worth an hour:
- [method shift, not benchmark bump — they replace X with Y]
- [claim is falsifiable: ablation in §4 isolates X vs not-X cleanly]
- [code + weights released, reproducible at home]
Where it might be wrong:
- [training distribution is narrow — open question whether the effect transfers]
- [comparison baseline is older than expected]
Read order:
1. §3 (method) — 8 min
2. §4 (ablations) — 12 min
3. §6 (limitations) — 5 min
Optional: §5 (extended experiments)
Empty days
If 0 papers survive the filter:
No surviving picks today. 14 papers filtered out: 6 benchmark-chasing, 5 incremental scaling, 3 position papers.
Worth scrolling yourself: huggingface.co/papers
Honesty over manufactured picks.
Rules
- One pick, not three. Three picks is a digest, and digests become wallpaper.
- "Where it might be wrong" is mandatory.
- Read order time-budgeted. Save the reader from skimming the whole PDF.
- Cite actual sections so the reader can verify the brief.
Related Skills
cyotee/bankr-communities
development
VerifiedTrustedCommunity
Bankr Space ↔ bankr.bot/agents two-way sync (BANKR-PROJECT-SYNC.md Paths B+C). Original tweets from GET /agent-profiles/:id/tweets shown on Spaces. Holder votes: yes/no or multiple-choice polls (1–24h) — HOLDER-VOTES.md (never say no poll feature). Agents: POST /api/agent/start-vote with symbol Space. Petition spaces: fee-right unit holders only. POIDH: create/list on bankr.space; fund/claim/vote on poidh.xyz. NOT Twitter audio Spaces.
SKILL.mdUpdated Jul 22, 2026
cyotee/b20-console
development
VerifiedTrustedCommunity
Inspect B20 token contract addresses on Base through B20 Console. Use this skill when a user asks whether a B20 contract is live, initialized, recognized by the B20 factory, what policies are attached, whether features are paused, or what risk flags are active. Calls the public B20 Console API and returns a concise risk summary with reasons. No API key required.
SKILL.mdUpdated Jul 22, 2026
cyotee/azzle
development
VerifiedTrustedCommunity
Post, claim, and settle agent tasks on the AZZLE protocol (Base mainnet, USDC escrow). Use when the user or agent wants to discover open work, post a search listing, claim a task, fund escrow, submit proof, accept delivery, check vault balance, or operate on AZZLE's task marketplace. Requires Bankr for swaps, approvals, and on-chain execution. NOT for modifying AZZLE smart contracts or running the azzle.org website.
SKILL.mdUpdated Jul 22, 2026
cyotee/autoboy
data-ai
VerifiedTrustedCommunity
AutoBoy by The Firm — the pre-launch order book for Bankr launches on Base. Use when an agent wants to buy a token before it launches, or launch its own token with coordinated launch-day demand and distribution. Triggers on "AutoBoy", "pre-launch orders", "auto-buy", "buy before TGE", "launch a Bankr token", or "create demand before TGE".
SKILL.mdUpdated Jul 22, 2026