CKBoost Post-Hackathon Testing & Feedback Report
CKBoost served as the live campaign management platform for the CKB AI Agent Hackathon, providing real testing data across both the campaign manager and user flows. This report highlights which issues were identified and fixed, which need to be addressed prior to completion, and enhancements that may be considered for future development. Also included is summarised feedback of using the platform by hackathon participants.
Overall perception
As a campaign manager, I found the process of creating campaigns relatively straightforward. All of the necessary fields were present, and a variety of options allowed me to customise the campaign appropriately. Setting and changing start/end dates for a campaign was smooth and immediately reflected on the campaign card. Extending or reopening a campaign required no additional steps. As creating a campaign can take some time, the campaign input page had a useful save feature which meant that I could return to the page at a later time and continue from where I left off.
Some of the UI around campaign and platform admin dashboards was a little confusing, but isn’t insurmountable and should be simple to improve. The main UX issue I found was around the costs associated with approving quests, especially multiple quests at once. This is mentioned in more detail below.
As a user, I found navigating the website quite easy. The initial page sync is quite long, and each page needs its own time to sync. Outside of this, I could explore and participate in campaigns, with an easy to follow user flow. Points were issued reliably, as were UDT/CKB. Tipping works as expected, although for future development it would be good to tip someone by their X or telegram handle which they could then claim on CKBoost. Telegram verification was surprisingly smooth, with a bot handling messages sent to my telegram account to confirm identity. I think user profile management and display could be improved, with did:ckb integration a logical step moving forward.
Users who participated in the hackathon were also asked to share their feedback.
Based on 10 respondents, overall reception to CKBoost was positive. Participants consistently praised the platform’s clean UI and ease of navigation, with one describing it as “clear and easy to use” and another highlighting that “simplicity stood out” as its most important quality. The gamified feel of the quest structure was also noted favourably, with the design described as looking “gamified” in a good way. Onboarding was flagged as straightforward by multiple participants, and the general flow of moving through quests was described as clear with no major issues encountered.
The most commonly reported friction points were slow page load times, issues with file uploads, and quest progress resetting on page refresh. A few users also noted that submission form inputs felt over-engineered for the information being collected, and that the leaderboard only showed the top 10 limited visibility for most participants. Detailed breakdowns of each of these are captured in the suggestions below.
Bugs Identified & Resolved
The following issues were identified and closed during the testing and campaign period:
Resolved before/during the campaign (no formal github issue)
- Quests not editable after creation. Quests could not be modified, added, or deleted once set. Resolved prior to campaign launch.
- Markdown not rendering correctly in long description. Formatting written in the campaign description editor was not being fully reflected when published. Resolved prior to campaign launch.
- Campaign commencement flow unclear. The path from campaign page to first quest was not intuitive. An additional button directing users straight to Quest 1 was added to address this.
- Campaign page intermittently failing to display long descriptions. When navigating directly to a campaign link, the long description would occasionally not appear on load. Resolved during the campaign period.
- Exporting quest data. Manually parsing through quest submissions on the campaign management side was flagged as time consuming. A feature was added to export quest submissions in XLS format.
Resolved via GitHub issues
- #41 Social interaction verification unreliable (tag-based search). Searching by tag to verify social interactions (e.g. posts, reposts) was not reliably picking up submissions. Resolved by switching to a replaceable event approach that directly stores the Nostr event ID for social interactions.
- #43 UI: Markdown formatting not fully rendering on publish. Campaign descriptions with formatted content (headers, bullet points, line breaks, text sizes) were only partially rendering after publishing, with bold being preserved but other styles lost. Resolved.
- #40 Expired event handling. Issues arising from expired campaign events were identified and resolved.
Issues needing resolution for M3
- #48 Reducing platform transaction costs. Some actions such as adding/removing quests and approving submissions incurred higher-than-expected CKB fees, while other operations had inconsistently low fees. Fee predictability and overall cost-to-interact must be addressed before mainnet.
- #49 Documentation searching error. The onboarding documentation currently does not display and the search bar returns an error. As this is part of the M3 deliverables, this needs resolution.
- Bug reports and “todo” items to clean up if addressed. Most of these are open issues from November 2025 and need to be formally concluded.
Enhancements for Next Development Round
Based on my testing and feedback from users, the following have been logged and remain open for future development. They are not strictly required for M3.
- Page load time. The initial sync takes 5-10 seconds to ensure the latest on-chain campaign data is visible. Several users flagged that this can feel like too long to wait. I would recommend exploring whether this wait can be reduced to improve UX.
- Submission form UX. The rich text/Markdown editor used for submission inputs (e.g. for a Telegram username or file upload) was flagged as unnecessarily complex. Simple text inputs and a standard file picker would better suit these fields.
- Pre-population of quest fields. Quest submission fields were auto-populating with data from previous quests, requiring users to manually clear them before entering new information. Other users reported quest details not saving on refresh, so more consistent behaviours around quest saving would be beneficial.
- Leaderboard pagination. The leaderboard currently displays only the top 10. Users suggested expanding this to at least 30–50 to give more participants visibility of their standing.
- Image upload limitations. From my testing, it appeared as though images above ~80kb were rejected by Nostr relayers or caused a long delay for confirmation. Other users reported a similar problem. A size warning or compression step would help. Alternatively, using a different method to handle larger files may be necessary
- Better identity integration. Tying wallet/address identity to quest submissions and the explorer was unclear on the campaign management side. Better profile display and identity context within the platform is needed. Additionally, better profile management and customisation options would give the site a more personalised feel.
- Export refinement. The export function exists but outputs need to be presented in a more user-friendly format.
A list of other potential feature additions is recorded on issue #31.
Conclusion
CKBoost has made significant progress since its last milestone, and its successful deployment as the campaign management platform for the CKB AI Agent Hackathon represents a meaningful checkpoint. The platform held up well under live conditions, with numerous bugs identified and resolved through the testing process. This is a huge credit to @Alive24 and his work on the platform. This is a credible platform to organise campaigns, drive engagement, and disburse rewards. I will now be looking into utilising this for monthly progress reporting of CKBuilders and strategies for social media campaigns.
Prior to final M3 completion and progression to mainnet, some work is needed around 1) transaction cost optimisation, 2) onboarding documentation and search functionality, and 3) concluding any outstanding github issues.