Second day of timeouts on the batch endpoint between 02:00 and 03:00. Same window both days, same error code.
ReproducibleFelo X Search on OpenClaw 2.0
OpenClaw 2.0
OpenClaw 2.0The channel hears about a problem before the dashboard does
Your customers complain on X, and someone in the team eventually spots it. Felo X Search watches the conversation and replies with the posts, the time, and the signal each one carries, so the alert arrives before the summary does.
- Install
- clawhub install felo-x-search
- Query
- Brand, product, or handle
- Comes back as
- Posts with a signal tag
- Reach
- Telegram, WhatsApp, Slack, Discord
Install from the registry
clawhub install felo-x-searchThis skill runs on
Read the OpenClaw 2.0 guideOpenClaw 2.0
Open-source personal assistant · release v2026.8.1
OpenClaw treats a skill as something with a lifecycle — install, review, then enable per agent — and answers wherever you already chat, so the reply has to hold up on its own.
- Registry
- ClawHub
- Review
- Skill Workshop keeps proposals and checks
- Reach
- WhatsApp, Telegram, Slack, Discord, native apps
Posts sorted by what they claim, so an alert is worth waking up for
A raw feed is noise with timestamps. The signal tag is what makes a chat alert actionable: a reproducible problem is a different message from a complaint, and both are different from an echo of your own announcement.
The status page still says all clear while the community reports otherwise. That gap is the story.
Status gapNo problems here, but we run the older SDK, so we are probably not a useful data point.
Scope limitAmplifying the vendor's announcement with no additional detail.
Echo, not reportingMake the alert rule match the signal, not the volume: one reproducible report is worth an interruption; ten echoes are not.
Three ways an alert channel goes wrong
All three are quiet failures, which is why they are worth writing down next to the enablement list.
A post is a lead, not a diagnosis
One reproducible report beats a hundred complaints, and the signal tag exists so that ranking survives into the alert rather than being flattened into a count.
Volume is not severity
An outage with four posts can matter more than a pricing argument with four hundred. Sort by what the post claims.
Polling burns a rate budget
A skill that queries every minute will spend the day rate-limited and report the quietest version of the day as if it were news.
Add it on OpenClaw 2.0Registry, review, then which agents may watch
A monitoring skill runs unattended, which makes the review step more important than the install step. Read what it queries and how often before you enable it in a channel.
From the official registry
clawhub install felo-x-searchSkill Workshop, before anything is live
- What the skill proposesThe query, the window, and the fact that this skill only reads. Nothing about it can post or reply, which is what makes it safe to leave running.
- What ran against itThe rate behaviour OpenClaw recorded, so a frequent schedule is a decision you make rather than a surprise you discover.
- What stays on recordApplied skills remain in the Workshop with their check history, which is the difference between an install you can audit and a folder somebody copied in.
Who gets it
Ops channel agentEnabledWhere an incident first becomes visible. Alert on reproducible reports and status gaps, not on mention volume.
Scheduled briefingsEnabledA weekly digest of what the market said, with the signals already sorted and the echoes dropped.
Main chat agentNot neededYou do not want this answering casually in your daily threads; a monitoring skill earns its place on a schedule.
Three checks before you trust the alert
Verification here is about specific posts, because a fabricated alert looks exactly like a real one.
Open one quoted post
Find it on X directly. A handle that does not resolve is the fastest way to catch a set that was assembled rather than collected.
Check the timing against the claim
The post should be recent relative to the incident, not merely recent relative to today. An old post in a fresh alert is how false incidents start.
Ask for the counter-example
Request the posts that contradict the summary. A set with none was probably filtered for agreement somewhere along the way.
Two prompts that produce a useful alert
The first is a tripwire, the second is a digest. Both belong on a schedule rather than in the moment.
Check X for reports of timeouts in the last six hours. If nothing is reproducible, reply nothing.What comes backSilence on a normal night, and a short list of reproducible reports when something is actually wrong.
Every Friday, summarise what people posted about our product this week. Group by signal and drop the echoes.What comes backA readable weekly digest that the team can act on rather than a wall of posts.
What OpenClaw users ask about Felo X Search
Can it alert in a specific channel?
The reply goes back to wherever the agent answers, and your approval rules decide what it may send. Point the scheduled task at the channel where incidents are already discussed.
Can it reply to posts on our behalf?
No. It reads. Nothing about it can write to an account, which is what makes it safe to enable on a monitoring agent.
How is this different from a social listening tool?
A listening tool gives you a dashboard and a volume chart. This gives you a classified set inside the conversation where you already respond to things.
What about sentiment?
Sentiment is a summary you cannot check. A signal is a claim you can open, quote, and argue with, which is more useful in a chat.
Let the channel hear about it first
Install from the registry, review it in Skill Workshop, and enable it on the agents that watch your product.
Every Felo skill for OpenClaw 2.0
Five skills written for OpenClaw 2.0. They follow the same pattern on this platform, so you can combine them in one session without re-explaining your setup.
Felo Search
Felo Slides
Felo YouTube Subtitling
Felo SuperAgentOne API key. A full agent toolkit.
Start with a skill, then compose the exact models, live data, knowledge, and output tools your OpenClaw 2.0 workflow needs — the same key every Felo skill already uses.
Models
Use cost-efficient LLM access when your agent needs custom reasoning or generation.
- LLM API
- Chat completions
- Model access
Live information
Give agents current web and social signals instead of relying on stale context.
- Web Search
- X Search
- Web Fetch
Knowledge and output
Ground work in reusable knowledge, then turn it into shareable deliverables.
- LiveDoc
- PPT generation
- Mindmaps
Creative tools
Extend the workflow from research to visual and office-ready results.
- Image generation
- Page tools
- Office outputs