~/project/.dsh/skills/ProjectLoads for this repository only. The right place to read a bundle before you trust it with the whole machine.
Felo Slides on dsh
DeepSeek Harness
DeepSeek HarnessTurn an investigation into a deck without leaving the session
The hard part of a review deck is not the design, it is remembering what you found three hours ago. Felo Slides builds the outline from the work the session just did, so the deck starts from evidence rather than a blank page.
Mount it
cp -r felo-skills/felo-slides .dsh/skills/This skill runs on
Read the DeepSeek Harness guideDeepSeek Harness
Open-source agent harness (developer preview)
dsh reads skills straight out of folders you point it at, so a Felo bundle is one copy away from being callable in a session — and one delete away from being gone.
- Loads from
- ~/project/.dsh/skills/ · ~/.agents/skills/
- Install is
- A folder copy, no build step
- Worth knowing
- Only direct children of a root are read
An outline you can argue with before it becomes slides
You get headings and one line of argument per slide. That is deliberately not a finished deck — it is the version you can cut, reorder, and disagree with while changes are still cheap.
Because the outline comes from the session, every claim on it can be traced back to the command or file that produced it.
Add it on dshSame one-folder install as any other dsh skill
Slide generation is a normal Felo bundle, so it mounts the same way. Nothing about deck output needs a different root or a longer setup.
~/.agents/skills/UserLoads in every project, and is the same folder other harnesses on this machine look at.
Mount the bundle
mkdir -p .dsh/skills
cp -r felo-skills/felo-slides .dsh/skills/Prove it loaded
List the skills you can call from this project.Ask for the deck at the end, not the beginning
The outline is only as good as the session behind it. These two prompts work because the work has already happened.
Turn the last hour of this session into a six-slide review: what we changed, what we measured, what we still do not know.What comes backAn outline whose slides map onto work that actually happened, so you spend your time editing rather than remembering.
Draft the same review for a non-engineering audience. Drop the implementation detail and keep the trade-offs.What comes backA second outline at a different altitude, which is usually faster than rewriting the first one by hand.
What the skill will not decide for you
Being clear about this is why the deck is usable. Everything the skill does not do is something you should keep owning.
It will not pick your argument
The outline reflects what the session found. If the session found the wrong thing, the deck is wrong in a tidy, well-spaced way.
It will not know your house style
Brand templates and slide masters are applied at export. Bring your own, or the deck will look generic in a way everyone recognises.
It will not cut for you
A six-slide deck and a twenty-slide deck are both one prompt away. Deciding which one the meeting needs is still your job.
Where the evidence for the deck comes from
Slides are the last step. These two are the ones that fill it in.
What dsh users ask about Felo Slides
Do I need a second API key for slides?
No. The same Felo key covers slide generation along with search, page fetching, and transcripts — that is the main reason to keep the CLI as the credential owner.
Can the session write the deck straight to a file?
The skill produces the outline and the export. Ask the session to save the outline to disk first, so you have a version to edit even if the export needs a second pass.
Is the outline stable if I run it twice?
Expect it to shift. The input is a conversation, not a template, so rerunning after more work gives a different deck — which is usually what you want.
Every Felo skill for DeepSeek Harness
Five skills written for DeepSeek Harness. They follow the same pattern on this platform, so you can combine them in one session without re-explaining your setup.
Felo Search
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 dsh 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