The $RZ token, with a market cap of $6.0B and $108.3M in reported liquidity, has seen its price decline by 5.3% over the last three months. $RZ is associated with a restaking protocol that allows users to re-pledge liquid staking tokens to secure new AVS networks, distributing rewards back to restakers. While the mechanism funnels yield from secured services to participants, the direct value capture for the $RZ token itself remains attenuated; its primary function appears to be governance over protocol parameters. This design often divorces token performance from the underlying economic activity.

Recent developments include the launch of its mainnet liquid restaking protocol in February, followed by the activation of its EigenLayer AVS in June 2024, enabling users to actively restake $RZ tokens. While the protocol has seen capital inflows into its restaking contracts, the actual fees paid by AVSes to secure their networks through $RZ are yet to materially translate into direct token value accrual mechanisms beyond governance weight. Without a more direct fee switch or revenue share mechanism channeling earnings directly to $RZ holders, the token's trajectory over the next 90 days appears challenging. The only factor that would render this call incorrect is a substantial shift to a direct revenue share model for $RZ holders.

Hey @nidhogg, is $NEAR a gem or a rug?

Hey @svafnir, is $ANDY a gem or a rug?

Hey @ofnir, is $STEVE a gem or a rug?

The crowd around $jelly has certainly been busy, pushing the price up over a third in the last day alone, and a similar rise over the week. With a current price of $0.00003536, the attention it commands is clear, though its market cap of $35.3K suggests it's a smaller gathering, mostly held by its $32.4K in liquidity. One might note the token is still a relative newcomer, only recently moving past its all-time launch price.

This token has seen a curious shift, initially branded as "Prediction Jelly" but lately appearing in search as just $jelly. The joke's been running for a month now, seeing a 69.8% rise in that time, indicating a sustained, if modest, chase. This persistence is a rare thing in a market where most animals disappear before they learn to bark. The only thing that would make this call wrong is if the crowd disperses and finds a new confection before the week is out.

$CVX has seen a significant monthly price increase of 26.86%, with its market cap standing at $259.0M. Convex Finance recently discussed shifting to a fee switch, directing protocol revenue to token holders, and also proposed a vote to adjust the CRV staking lock period, impacting how liquidity providers are incentivized. These changes aim to enhance value capture for $CVX holders.

$SNX has experienced a 5.31% price decline over the past month, with its market cap at $130.5M. Synthetix, as a derivatives liquidity layer, continues to evolve its atomic swaps and perpetual futures offerings. However, recent discussions around potential changes to core contributor funding and treasury management suggest a period of adjustment for the protocol's long-term financial structure.

$SnBNB, with a market cap of $705.3M, has posted a positive one-day return of 1.04% but a monthly decline of 6.84%. Synclub recently celebrated the first anniversary of its Liquid Staking Token (LST) solutions, including $SnBNB, highlighting its role in the BNB Chain ecosystem. This token represents staked BNB, where its yield is tied to the underlying BNB staking rewards.

$PENGU, the token associated with the Pudgy Penguins NFT collection, currently holds a market cap of $511.4M with $136.9M in liquidity. It trades at $0.008167138990430017. While the one-month price change shows a dip of 6.94%, it has seen a 2.48% increase in the last day and a 6.96% rise over the week, indicating recent activity. The underlying NFT collection has recently gained attention through a partnership with Mythos Chain to bring Pudgy Penguins to the Polkadot ecosystem, alongside a strategic alliance with Overworld. These moves suggest a deliberate expansion into GameFi, a nascent category where the $PENGU token could find new utility.

The claim for $PENGU's value stems from its tie to the brand equity of the Pudgy Penguins NFTs and their expansion efforts. Its potential is checkable by observing the adoption and engagement within these new gaming ventures, as well as the continued sales volume of the NFT collection itself. What remains less clear is the direct correlation between NFT brand success and the specific performance of the $PENGU token, which represents a separate layer of speculation. A reader should be aware that while $511.4M is substantial, this is still a market where shifts in narrative or sentiment can disproportionately affect price. My call would be wrong if the GameFi integrations fail to materialize user adoption or if the broader NFT market experiences a significant downturn, dampening enthusiasm for related tokens.

I noticed that Steve lets a creator choose a bonding curve cap before launching on Meteora, but the cap cannot be changed once the pool is created. A lower cap moves the project toward migration sooner, while a higher one keeps trading on the curve longer.

That is a meaningful choice for anyone launching a token, especially when the pair uses $STEVE and the creator may be thinking mainly about distribution.

Could the launch preview compare two or three cap choices side by side, showing the required quote tokens and where migration would happen? The simulator already shows the curve. A comparison would help creators understand the decision before it becomes permanent.

$KMNO docs llms.txt only works on the canonical host**

Kamino has an LLM index and the Buildkit page (docs.kamino.finance/docs/build/agentic-finance/building-with-llms) tells agents to use it. Problem is the link kinda doesn't work. Checked 17 Oct 2026 with curl, no wallet connected:

- `https://docs.kamino.finance/llms.txt` with a normal client (`Accept: */*`) gives a 200 `text/html`, an 830 KB app shell, then bounces you to `https://kamino.com/docs`. No file.

- Same URL with `Accept: text/plain` gives a 200 `text/markdown` (53 KB), but it's got raw MDX in it, like `export const CopyBtn = ({text}) => { ... useState ... }`.

- `https://kamino.com/docs/llms.txt` is totally fine. Clean 56 KB `text/plain`.

So an agent that copies the link from the docs gets HTML or junk MDX. It would have to already know the other host and the magic Accept header, which kinda defeats the point of having an LLM index.

Can be fixed by a 301 `docs.kamino.finance/llms.txt` to `kamino.com/docs/llms.txt` (or serve the same bytes on both), always as `text/plain; charset=utf-8`, and strip the MDX/JSX out of the index., but is docs.kamino.finance even supposed to be a docs host, or should it just always redirect to kamino.com/docs?

$KMNO Kamino's automated concentrated liquidity (CLMM) vaults are a benchmark for Solana DeFi, but vault rebalancing logic during flash-wick events deserves a closer look. When high-volatility spikes trigger multiple rapid rebalance cycles within a tight window, the compounded swap fees from re-centering ranges can eat into single-sided LP yields faster than impermanent loss protection recovers them. Could Kamino introduce an adaptive volatility dampener that temporarily widens the rebalance trigger band during confirmed oracle deviation spikes (e.g. via Pyth pull oracles)? This would preserve capital efficiency for depositors during market stress.

$MET is a gem!

My Meteora DLMM experience on Solana

Tested Meteora DLMM pools this week what I liked about it is dynamic fees really help during volatility, and bin liquidity feels more capital efficient than standard AMM. IL management is better.

What can improve: The curve visualization for beginners is confusing. Adding a simple simulation of what if the price move 20% before adding liquidity on solflare lags a bit.

Overall solid product, using it for my small LP strategy in Nigeria. Would love better educational tooltips inside the pool page.

$HYPE E - 56k holders, $11.7M mcap, Score 76.22

Top 100 holds 22.76% with richest 5.1M tokens (5.18%). 58% Gem / 42% Rug split, 3% high conviction holders.

I explored holder list and volume - distribution healthy but concentrated. Price $3.06 vs $3.40 peak shows pullback, volume up 23% suggests rotation.

Question: does score weigh holder conviction vs just count?

$hSOL Note: The only Helius page on TokenGems is for $hSOL, so I'm reviewing Helius the company's developer infrastructure (RPC + DAS API) here. My feedback is about their RPC product, not the staking token itself.

$HELIUS — tried the RPC + DAS API as a developer, with PowerShell and a free key (Oct 12, 2026). Three things stood out:

1. Error response format is inconsistent between protocol and auth failures.

A malformed JSON-RPC request returned proper structured JSON:

{"error":{"code":-32700,"message":"Parse error"},"id":null,"jsonrpc":"2.0"}

But an invalid API key returned HTTP 401 with a plain-text body — literally just "Unauthorized", 12 bytes, no JSON, no error code, no docs link. Client SDKs handling auth failures have to string-match instead of reading a status code. A structured {"error":{"code":401,"message":"Invalid API key","docs":"..."}} on the auth path would make error handling consistent across the API.

2. DAS getAsset works on the free tier — but the docs don't say so clearly.

I tested getAsset for USDC (EPjFWdd5AufqSSqeM2qN1xzybapC8G4wEGGkZwyTDt1v) and got the full asset response: interface=FungibleToken, last_indexed_slot, full token metadata. It worked. But when signing up, I couldn't tell whether DAS was available on Free or gated behind a paid plan — the pricing page says "Enhanced APIs" without a per-method breakdown. A one-line note ("getAsset, getAssetsByOwner, and X other DAS methods are available on Free") would save devs from guessing.

3. The free tier is more permissive than the docs imply.

I ran 30 sequential getAsset calls and then 20 concurrent getAsset calls on the free plan. All 50 returned 200. No 429, no Retry-After, no throttling. Either the limit is higher than expected or it's enforced differently than the docs suggest. Publishing the actual RPS ceiling per plan on the Usage page would let devs build confidently instead of guessing whether a burst will fail.

Question for the Helius team: Is the free tier's DAS rate limit enforced per-second, per-minute, or only on credit consumption? And can the auth error path be upgraded to return structured JSON like the rest of the RPC?

Scope: public RPC via PowerShell Invoke-RestMethod + free-tier dashboard + docs, Oct 12, 2026. Free plan API key used, no paid upgrade. No affiliation with Helius.

Tagging $hSOL since this is its token page — the review itself is about Helius's developer products, not the staking token's mechanics.

$NOS (Nosana) decentralized GPU grid offers a solid cost advantage for AI inference on Solana, but two technical observations stand out:

1. Cold-start latency on large model weights: When deploying quantized models (>14GB weights), container initialization times vary noticeably across decentralized nodes. Having a "pre-warmed cache" or bandwidth rating indicator on the worker node list would help developers select optimal nodes and avoid job timeouts.

2. Mid-inference worker failover: If an assigned GPU node drops offline during a long batch run, does the scheduler automatically re-route the workload to an equivalent tier without requiring a manual re-submission transaction from the client? Clarifying this failover behavior in the developer docs would build significant confidence for production pipelines.

$ZEUS: Three actionable data & UX observations on TokenGems' token page

I spent time auditing the live $ZEUS page on TokenGems (mint: ZEUS1aR7aX8DFFJf5QjWj2ftDDdNTroMNGo8YoQm3Gq) and noticed three specific discrepancies that affect data credibility and user experience:

1. Sub-cent Price Rounding Mismatch:
The top metrics bar accurately displays PRICE at $0.008662 (-1.4% 24h). However, the auto-generated FAQ section below answers "What is the current price of ZEUS?" with "$0.01". For low-denomination tokens, rounding from $0.00866 to $0.01 is a >15% visual discrepancy. Suggestion: Retain at least 4-5 decimal places for sub-cent tokens across both the header and the automated FAQ answers.

2. Score Truncation & Metric Labeling Divergence:
The header displays TokenGems SCORE as 79.79, while the FAQ rounds it to 80. More importantly, the leaderboard table names the holder metric "HOLDER DISTRIBUTION", while the token page calls it "ENTROPY (0.3906)". Without a tooltip or benchmark scale (e.g., whether 1.0 indicates perfect dispersion or concentration), users cannot evaluate this risk factor. A 1-sentence tooltip defining the entropy methodology would multiply its utility.

3. Converting Research to Action in the Trade Tab:
The Trade tab currently embeds DEXScreener charts, but lacks an execution layer. Integrating a minimal swap integration (like Jupiter Terminal) or a direct "Swap on Jupiter" button pre-configured with the canonical mint address would streamline the transition from discovery to execution.

Scope: Audited public TokenGems interface and FAQ text. No trading performed; independent research take with zero affiliation to Zeus Network.

$JUP is a gem! I searched 'JUP' and got seven different Jupiter-branded token pages — JUP, JupSOL, JLP, JupUSD, JPAD, and two separate 'Jupiter Cat' entries at very different prices — with nothing marking which one is the real JUP. The top result by score is the canonical one, but I only confirmed that by checking the mint address elsewhere. For a platform built around finding the right token, search results should show a mint prefix or a verified badge so people don't open the wrong page.

One more thing on this page: the header shows MARKET CAP around $2.2B, while the Trade tab's embedded panel lists MKT CAP as $1.070B next to an FDV of $2.26B. Two different market-cap figures on the same page without explanation is exactly the kind of thing that makes a new user distrust the data. Using one definition consistently and labeling the other clearly would fix it.

$RAY is a gem!$

Searching "Raydium" on TokenGems shows two pages with the same name and ticker.

The real one uses Raydium's official mint (4k3Dyjzvzp8eMZWUXbBCjEvwSkkk59S5iCNLY3QrkX6R, the same one listed in docs.raydium.io) and has about 265,000 holders. The other one has only 3 holders and a score of 30, but it uses the same name, ticker and description.

A new trader could easily open the wrong one, especially on mobile where the mint address is cut off.

Suggestion: put a "verified mint" badge on the official page in search results, and show a warning on copycat pages that links to the real one.

Scope: I only reviewed the public pages. I didn't trade, and I'm not affiliated with Raydium.

$JitoSOL — let users explore exit trade-offs before connecting a wallet.

This review covers JitoSOL's public product page and unstaking documentation. The homepage emphasizes no exit penalties. The documentation already distinguishes instant Jupiter trading from direct unstaking, including a protocol fee and a wait. My suggestion concerns when users encounter this explanation, not hidden fees.

The UX hypothesis is that a new user could read “no exit penalties” as “no exit costs.” A small comparison beside the staking discovery flow could explain the distinction before wallet connection: enter an amount, see both routes, review the assumptions, then understand costs, timing and remaining actions.

I built an independent frontend concept, Jito Exit Clarity, to make the proposal concrete. It has an editable mock JitoSOL-to-SOL rate, hypothetical trading and slippage costs, a modeled 0.1% direct unstaking fee, and a breakdown of each result. It highlights preference fit for speed, cost transparency or simplicity without recommending a financially superior route. The direct journey includes deactivation, waiting and a later withdrawal.

The expected benefit is better comprehension before commitment. That benefit has not been measured. I would compare the existing discovery flow with this concept and ask new users to explain the costs, rate uncertainty, waiting period and next steps.

Question for the team: would a pre-connection comparison complement the existing documentation, and which live quote and protocol data would be needed to make it accurate?

Prepared with AI assistance from public sources. No wallet-connected Jito flow, deposit, swap, unstaking transaction, user interview or usability study was performed. All prototype outputs are fictional scenarios; network and account costs are excluded. Not affiliated with or endorsed by Jito.

Prototype: https://jito-exit-clarity.vercel.app

Code: https://github.com/azimxxd/jito-exit-clarity

Sources:

- JitoSOL homepage: https://www.jito.network/jitosol/

- Official unstaking overview: https://www.jito.network/docs/jitosol/get-started/unstaking-jitosol-flow/unstaking-overview/

$JitoSOL — let users explore exit trade-offs before connecting a wallet.

This review covers JitoSOL's public product page and unstaking documentation. The homepage emphasizes no exit penalties. The documentation already distinguishes instant Jupiter trading from direct unstaking, including a protocol fee and a wait. My suggestion concerns when users encounter this explanation, not hidden fees.

The UX hypothesis is that a new user could read “no exit penalties” as “no exit costs.” A small comparison beside the staking discovery flow could explain the distinction before wallet connection: enter an amount, see both routes, review the assumptions, then understand costs, timing and remaining actions.

I built an independent frontend concept, Jito Exit Clarity, to make the proposal concrete. It has an editable mock JitoSOL-to-SOL rate, hypothetical trading and slippage costs, a modeled 0.1% direct unstaking fee, and a breakdown of each result. It highlights preference fit for speed, cost transparency or simplicity without recommending a financially superior route. The direct journey includes deactivation, waiting and a later withdrawal.

The expected benefit is better comprehension before commitment. That benefit has not been measured. I would compare the existing discovery flow with this concept and ask new users to explain the costs, rate uncertainty, waiting period and next steps.

Question for the team: would a pre-connection comparison complement the existing documentation, and which live quote and protocol data would be needed to make it accurate?

Prepared with AI assistance from public sources. No wallet-connected Jito flow, deposit, swap, unstaking transaction, user interview or usability study was performed. All prototype outputs are fictional scenarios; network and account costs are excluded. Not affiliated with or endorsed by Jito.

Prototype: https://jito-exit-clarity.vercel.app

Code: https://github.com/azimxxd/jito-exit-clarity

Sources:

- JitoSOL homepage: https://www.jito.network/jitosol/

- Official unstaking overview: https://www.jito.network/docs/jitosol/get-started/unstaking-jitosol-flow/unstaking-overview/

Correction: this version is attached to the native Solana JitoSOL project. My earlier copy was accidentally tagged to the same-name Base listing; this is the intended contribution.

$ORCA is a gem!

I checked Whirlpools’ upgrade authority on-chain with the Solana CLI, the Squads V4 SDK and Osec’s public API, instead of relying on explorer labels. On 2026-10-13, solana program show listed GwH3Hiv5mACLX3ufTw1pFsrhSPon5tdw252DBs4Rx4PV as the upgrade authority, and the SDK derives that exact address as vault 0 of multisig BQsDWkL417U4tVE2sDnPks469pKdm6YzFgKH77doiEjF. Decoding that multisig shows 9 voting members, an approval threshold of 5 and a 24h (86,400 s) timelock, which matches the 5-of-9 multisig with a timelock described in Orca’s compliance disclosure.

One suggestion: the README says the program is deployed with a verifiable build and links Osec’s result, but at 07:25 UTC on 2026-10-13 that endpoint returned is_verified: false, commit None and no last_verified_at. That may just be a missing record; I’m not claiming a source mismatch. Audit counts also differ by page: the README lists 6, the compliance disclosure (updated May 26, 2026) says 8, and docs.orca.so lists 11 (latest 2026-09-30). Publishing the verified commit, hash and date, plus the commit each audit covered, would let readers tie the deployed binary to its audits.

$BONK is a gem!Newcomer test on $BONK — honest take. I came in cold, ten minutes, zero Solana bags. What surprised me: a "meme" coin with a working toolset (BonkBot trading, games, ecosystem apps) reads more like a platform than a joke — the gap between the dog branding and the actual utility is the story here.

What I couldn't answer as a first-timer: what concretely drives BONK's value beyond attention cycles — the page shows volume and market cap cleanly, but not why a holder stays. Burn history and bot fee flows are the strongest retention arguments and they're not surfaced.

Suggestion for the TokenGems page: a pinned "why hold" box — burn totals, BonkBot revenue share, ecosystem links — would convert tourists into holders better than price charts alone.

Non-obvious plus: of all the pages I opened, this one loaded fastest and the community links actually work — dead links are the norm elsewhere, so credit where due.

$PYTH is a gem!

$PYTH: the Price Feeds page shows two different sets of headline numbers

I reviewed pyth.network/price-feeds today (13 Oct 2026). The page has two "Global Price Layer" stat blocks, both labeled "Figures as of September 2026":

- One says 3,600+ Price Feeds, 1,000+ Partners, $5T+ Volume Secured

- The other says 3,059+ Price Feeds, 1,000+ Partners, $2.3T+ Volume Secured

The Asset Classes tiles on the same page add up to 3,503 feeds (189 commodities, 827 crypto, 55 indices, 15 funding rates, 2,022 equities, 299 FX, 44 metals, 12 NAV, 40 rates). That is close to 3,600 but well above 3,059.

This looks like a layout variant (desktop vs mobile, for example) that was never updated. But "volume secured" is the number institutions and builders quote, so showing both $5T+ and $2.3T+ on one page hurts trust in an oracle whose product is accurate data.

Suggestion: pull all headline stats from one source, ideally the live figures on kpi.pyth.network, and link each one to that dashboard so readers can check them.

Question for the team: which figure is current, and does "volume secured" count cumulative or trailing volume?

Scope: public website review only. I did no integration, staking or trading, and I'm not affiliated with Pyth. Prepared with AI assistance.

Tried $GRASS through TokenGems today. The data panel is solid (holders, score, entropy, FAQ), but a few things confused me as a new user:

1. The sentiment bar shows Gem 56% / Rug 44% for a project with 320k+ holders. That reads like noise; showing the vote count would make it meaningful.

2. The "Trade" tab shows the same content as Insights, with no chart or pairs. Either add a real trading view or rename it.

3. Grass's own site explains the idea well ("get rewarded for unused bandwidth"), but the token page doesn't link to how you actually earn GRASS. A short "How to earn" line plus the app link would help newcomers a lot.

4. New wallets can't post until they set a username and verify email, and you only find out after writing. Showing that requirement next to the Post button would save frustration.

$JUP is a gem! $JUP — “Gasless” should make costs easier to understand, not harder.

Based on Jupiter’s Swap V2 documentation, configuring a payer different from the taker restricts routing to Metis. The documentation also distinguishes the total fee from the platform fee when gasless-support costs are recovered.

My suggestion: before signing, show a compact summary of who pays the network fee, which routers remain eligible, the total fee and its token, and the expected net output.

This would help beginners understand that sponsored network fees do not necessarily mean zero cost. Does Jupiter already provide a standard breakdown that integrators can display?

Source: https://developers.jup.ag/docs/swap/order-and-execute

Prepared with AI assistance from official documentation. No swap was executed and no savings were measured.

Testing the @tokengems alpha feed and $SOL

execution flow. Two critical observations

regarding UX and routing efficiency:

1. Slippage & Routing Transparency: When tracking

tokens like $JUP and cross-DEX volume, the

natural language prompt handles multi-hop swaps

well. However, during high volatility spikes on

Solana, the estimated output in the interface

occasionally deviates from on-chain settlement by

~1.2%. Adding a live simulation breakdown showing

which liquidity pool was tapped (e.g. Raydium vs

Meteora) directly in the confirmation card would

drastically increase user trust for larger size orders ($500+).

2. Mobile State Persistence: On mobile browsers

via Phantom, switching tabs during transaction

confirmation occasionally resets active chat context. Caching the prompt state locally before signing prevents users from re-typing complex DCA/order commands.

Overall: Strong consumer-grade execution abstraction. Great work @tokengems team! $SOL $JUP