DAG Skill Platform
Skill creation w NLP/template, run interactive/adhoc tasks
9 followers
Skill creation w NLP/template, run interactive/adhoc tasks
9 followers
Team/department level low-barrier enterprise Skill creation (with template or natural language) for all users—turn expertise/SOPs into IP, third-party skiill/MCP server installs, auto-generate basic MCPs. DAG-powered orchestration automates workflows with real-time monitoring, interactive skill for comprehensive tasks such as RFP/technical document/finacial consulting etc., native Microsoft integration, scheduled task automation, knowledge base &search engine MCPs adds information on demand.
Free
Launch tags:Artificial Intelligence
Launch Team
Render: Get up to $100K in startup creditsCloud infrastructure for founders who want to move fast and NOT break things
Promoted
1.Save enterprise expertise and experience to skill as asset is not a AI expert/programer's job now with this DAG skill platform.
2.Interactive skill put human in the loop for high quality professional senarios.
3.Light weight knowledge base indexed documents for retrieval/conversion/generation.
4.Timer put natural lanaguage task executable.
5.Microsoft login for user segregation with own generated/installed skill/MCP but with common skill/MCP for all, so this platform suitable for team/department.
@kymetmadenufbg At this moment the DAG is orchestrated behind, only error happens and retrying failes it will show messages on UI, and can retry manually. At this moment 'success silient' fits. Of course each stage of Interactive skill execution will show result (even partial missing) so still have chance request it redo/add something. Good idea, will find a good senario.
Please click Help on upper right for more case demos.
Turned a messy onboarding SOP into a working skill in about ten minutes using the template flow, and the DAG orchestration actually made sense for once. Curious how it handles edge cases when a step fails halfway through.
@enaypbab You can input 'failed task' then it will list failed task list, find the task number and input 'retry task task_id', it will try to run from there, as each failed flow has snapshot.
The Microsoft integration is a nice touch but a Slack or Teams-style notification when a DAG workflow fails mid-run would save so much debugging time, maybe even with a quick retry button right in the chat. That would make the monitoring feel less like something you have to babysit and more like something that actually helps you.
@aysel1129285 Good idea. I usually set adhoc task and request send the output to mail as usually execution time is long depend on complexity. Failed task now can retry for a few times, if still fails the Failed Task list is available for manual retry. Interactive creation needs in time input/clarification/request redo, that is OK.Thanks
Have you thought about adding a visual skill dependency map? When chaining a bunch of skills and MCPs together through DAG orchestration, it gets hard to keep track of which skills feed into which. A drag-and-drop graph view showing inputs, outputs, and execution paths would make debugging failed workflows way less painful.
@kymetmadenufbg At this moment the DAG is orchestrated behind, only error happens and retrying failes it will show messages on UI, and can retry manually. At this moment 'success silient' fits. Of course each stage of Interactive skill execution will show result (even partial missing) so still have chance request it redo/add something. Good idea, will find a good senario.