Any change to the northern line service today? Send the link you used.What comes backA short reply with the current status, the source, and the time it was published — not a summary you have to take on faith.
Felo Search on OpenClaw 2.0
OpenClaw 2.0
OpenClaw 2.0Ask in the chat you already use, and get an answer you can check
OpenClaw meets you where you talk, and a chat reply is easy to believe. Felo Search is what stops it being a confident guess: the answer arrives with the page and the date behind each claim, in the same thread you asked in.
Install from the registry
clawhub install felo-searchFelo Search on OpenClaw 2.0
- Installclawhub install felo-search
- Comes back asClaim, source, date
- ReviewSkill Workshop keeps the record
- ReachEvery channel your Claw answers on
This 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
An answer with the page and the date behind every line
In a chat, the reply is the deliverable — there is no appendix to check later. So the sourcing has to be in the message: each claim carries its source, and anything unconfirmed says so instead of blending in.
ClaimSourcePublished
The strike covers the northern line onlyoperator status page4 hours ago
Service resumes Saturday morningoperator press releaseYesterday
Freight is unaffected so farlogistics trade press2 days ago
The union meets again on Tuesdaynational news wireYesterday
The port schedule is unchangedport authority notice5 days ago
Ask for the sources in the message itself: a link and a date per claim, and a visible note when something could not be confirmed. A reply that says "unconfirmed" is more useful than one that does not mention it.
Add it on OpenClaw 2.0Install from ClawHub, review it, then enable it where you want it
OpenClaw 2.0 treats a skill as something with a lifecycle rather than a file that appears. Installing it and enabling it are two different steps, and the second one is per agent.
From the official registry
clawhub install felo-search
# or run the Felo installer script for all five at onceSkill Workshop, before anything is live
- What the skill proposesThe skill body, the files it adds, and the command it calls. Read it here before it is applied, because a skill is a procedure the agent will follow.
- What ran against itValidation results and whatever checks OpenClaw recorded for this proposal, so the applied list is a set of decisions rather than a set of accidents.
- 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
Main chat agentEnabledYour daily threads are exactly where an unsourced reply is most tempting and hardest to notice.
Scheduled briefingsEnabledThe recurring morning summary benefits most: dated sources turn a digest into something you can forward.
Team channelReview firstLeave this off until you have read a few answers yourself. A group thread has more eyes and less patience for a wrong figure.
Two questions worth asking in the thread
Both produce a message you can act on straight away, with the evidence already inside it.
Check whether the port schedule changed this week and reply only if it did.What comes backEither a sourced change or silence, which is the useful behaviour in a chat you leave running.
Three checks after the install
The Workshop shows that a skill was applied; only a live question shows that it is being called.
It appears in Skill Workshop
With its check history, not just as an installed line. That record is what you audit later.
It is enabled for the right agent
A skill installed for every agent when one needed it is a wider change than you intended. Check the per-agent list.
A live answer carries a date
Ask something recent and look for the publication time. A fluent answer with no dates means the skill never ran.
Three things a chat reply should not hide
The format is quick and ephemeral, which makes these worth stating rather than assuming.
Confidence is not evidence
A fluent reply in a chat reads as authority. The link and the date are what let you tell a live result from a remembered one, so they belong in the message rather than behind it.
A key on the machine is still a key
The skill calls the Felo CLI, so the credential sits in your user config. That is the safer arrangement than an inline key, but it is still worth knowing where it lives.
Group channels change the stakes
An unsourced answer in a direct message costs you a minute. The same answer in a team channel costs somebody else an afternoon.
What OpenClaw users ask about Felo Search
Does installing it change anything by itself?
No. Installing puts the skill in the Workshop with its checks; enabling decides which agents may call it. Keeping those two apart is the point of the 2.0 pipeline.
Which channels does it work in?
Wherever your assistant already answers — Telegram, WhatsApp, Slack, Discord, or the native apps. The skill runs on the agent, so the channel is only where the reply lands.
Can a scheduled job use it?
Yes, and that is the strongest use. Scheduled work is consistent across the agent, the Control UI, and the CLI in 2.0, so the same sourced answer can arrive without you asking.
What if the answer cannot be verified?
A good run says so. Ask for the phrase "unconfirmed" to be used literally, and a hedged reply becomes a signal rather than noise.
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 Slides
Felo YouTube Subtitling
Felo X Search
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