Search filter nonSuspicious=true doesn't hide certain skills

[rich-song]

Bug Report

The Clawhub search with the nonSuspicious = true filter doesn't hide several skills created by @steipete

Affected Skills

The following skills are still appearing despite the filter:

• https://clawhub.ai/steipete/gog
• https://clawhub.ai/steipete/sonoscli
• https://clawhub.ai/steipete/notion
• https://clawhub.ai/steipete/nano-banana-pro

Expected Behavior

Skills flagged as suspicious should be hidden when nonSuspicious = true filter is applied.

Actual Behavior

The skills listed above are still visible in search results despite the filter being enabled.

[steipete]

they are false positive and since I also build the thing you can trust me.

…

On Fri, Feb 27, 2026 at 12:33 AM Richard Song ***@***.***> wrote: *rich-song* created an issue (openclaw/clawhub#538) <#538> Bug Report The Clawhub search with the nonSuspicious = true filter doesn't hide several skills created by @steipete <https://github.com/steipete> Affected Skills The following skills are still appearing despite the filter: - https://clawhub.ai/steipete/gog - https://clawhub.ai/steipete/sonoscli - https://clawhub.ai/steipete/notion - https://clawhub.ai/steipete/nano-banana-pro Expected Behavior Skills flagged as suspicious should be hidden when nonSuspicious = true filter is applied. Actual Behavior The skills listed above are still visible in search results despite the filter being enabled. Screenshot.2026-02-26.at.3.29.59.PM.png (view on web) <https://github.com/user-attachments/assets/e977829d-9f8b-4840-a84c-e3c0aeb5964d> — Reply to this email directly, view it on GitHub <#538>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAOI7LOKPJ4T4CL2WURHOL4N57EDAVCNFSM6AAAAACWBEIUZKVHI2DSMVQWIX3LMV43ASLTON2WKOZTHE4TQMRXGYYTCOI> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[rich-song]

I trust you but for consistent standards, should they be marked as exceptions somehow? Maybe official tag like https://clawhub.ai/steipete/mcporter ?

[steipete]

Nah, just a temporary workaround since the other way caused me even more emails and tickets

…

On Fri, Feb 27, 2026 at 12:47 AM Richard Song ***@***.***> wrote: *rich-song* left a comment (openclaw/clawhub#538) <#538 (comment)> I trust you but for consistent standards, should they be marked as exceptions somehow? Maybe official tag like https://clawhub.ai/steipete/mcporter ? — Reply to this email directly, view it on GitHub <#538 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAOI7LCULKGPZOFDBXK4QT4N6AXLAVCNFSM6AAAAACWBEIUZKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTSNRZHA4TQNRVGY> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[rich-song]

The following skills are appearing despite the filter:

• https://clawhub.ai/JimLiuxinghai/find-skills
• https://clawhub.ai/halthelobster/proactive-agent

[steipete]

Active implementation exists in PR #537, which adds nonSuspiciousOnly filtering support and related regression coverage. Keeping this issue open as the tracking issue for that PR rather than closing it as a duplicate.

[steipete]

Fixed on main.

Relevant changes:

• fc22bfb keeps paginated nonSuspiciousOnly browse results on the correct index path during backfill fallback.
• a45c5b9 stabilizes browse pagination during safety backfill and adds regression coverage in convex/skills.listPublicPageV2.test.ts.
• Current list filtering on main applies highlightedOnly and nonSuspiciousOnly together and skips fully filtered pages instead of leaking suspicious entries.

I re-ran bun x vitest run convex/skills.listPublicPageV2.test.ts on March 8, 2026 and it passes.

Closing this since the non-suspicious browse filter behavior is fixed on main. If you still see suspicious entries after the next deploy, comment with the exact query URL and affected slugs.

[rich-song]

Thanks for the updates. I see that the fix correctly filters out this skill (https://clawhub.ai/TheSethRose/agent-browser) but the following skills still appear using this query URL (https://clawhub.ai/skills?sort=downloads&nonSuspicious=true). Also, it looks like skill flagged banner appears when VirusTotal marks it suspicious and not when OpenClaw scan does. Happy to take a stab as well.

• https://clawhub.ai/JimLiuxinghai/find-skills
• https://clawhub.ai/halthelobster/proactive-agent