You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Curious what the community and contributors think about this.
This repo is a pretty interesting experiment in what open source can do at this scale. Hundreds of specialized agents spanning everything from civil engineering to FedRAMP compliance to Blender addons, built and refined by a community rather than a single team of domain experts. That’s a genuinely new kind of thing, and it’s worth sitting with what it means and the potential impact something like this can have.
One question that came to mind browsing the roster: right now there’s no visible information about how any given agent’s expertise was developed, no author or reviewer attribution, no note on sourcing (practitioner experience, compiled standards docs, community iteration, etc.), nothing at the individual file level. That’s not unusual for open source generally, plenty of great projects are pseudonymous or lightly documented, but a Blender addon agent and a FedRAMP or legal document review agent probably don’t carry the same weight if the guidance is off. One gets you a broken script, the other could look plausible while being wrong in ways that surface much later.
A few things worth hearing people's thoughts on to get the conversation going:
1. A short “how these agents are built” note somewhere in the README. Even a general statement about process would help people have a better understanding.
2. Optional frontmatter fields on agent files (source, reviewed_by, confidence or similar) so provenance is visible per-agent.
3. Segmenting by risk. Given the size of the roster, asking for this at the level of every agent probably isn’t realistic. It might make more sense to focus on the domains where being wrong has real consequences (compliance, legal, medical, security, finance) and leave the rest as-is.
Happy to help draft something concrete to brainstorm on more specifics (README section, frontmatter convention, whatever direction people land on) if it’d be useful. Curious to hear what others think. Thanks. Hank
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
Curious what the community and contributors think about this.
This repo is a pretty interesting experiment in what open source can do at this scale. Hundreds of specialized agents spanning everything from civil engineering to FedRAMP compliance to Blender addons, built and refined by a community rather than a single team of domain experts. That’s a genuinely new kind of thing, and it’s worth sitting with what it means and the potential impact something like this can have.
One question that came to mind browsing the roster: right now there’s no visible information about how any given agent’s expertise was developed, no author or reviewer attribution, no note on sourcing (practitioner experience, compiled standards docs, community iteration, etc.), nothing at the individual file level. That’s not unusual for open source generally, plenty of great projects are pseudonymous or lightly documented, but a Blender addon agent and a FedRAMP or legal document review agent probably don’t carry the same weight if the guidance is off. One gets you a broken script, the other could look plausible while being wrong in ways that surface much later.
A few things worth hearing people's thoughts on to get the conversation going:
Happy to help draft something concrete to brainstorm on more specifics (README section, frontmatter convention, whatever direction people land on) if it’d be useful. Curious to hear what others think. Thanks. Hank
All reactions