Feature Description
I would like the option for lola to not update the AGENTS.md with transient text (installed skill descriptions for example)
Problem or Use Case
As a project maintainer, I use lola sync and .lola-req to recommend AI skill modules to other developers. I also use an AGENTS.md file in the repo to provide project specific context and rules to any agent developers are using.
When I run lola sync, it injects the installed skills and some other information into the AGENTS.md. This creates some issues:
- Since I have the file checked in, it now reports changes when skills change, or when I use a different version of a skill than someone else.
- If a user does not use lola, the AGENTS.md file will be wrong and could cause issues.
Category
None
Proposed Solution
I like that and understand why lola injects the information into the agent context. However, I think it would be better if lola wrote that information to a different file, maybe skills/lola.md or something, and then updated the AGENTS.md with a simple redirect "Check for lola.md and read it if its present"
This option could be implemented as a cli arg. I also think it would be nice to have it in a config file (or .lola-req although im not sure it fits.) because that type of setting would be best applied project wide and not at the user level
Alternatives Considered
No response
Complexity Estimate
I don't know
Additional Context
I really only use lola sync, so I am unsure if this problem extends to other cli operations
Feature Description
I would like the option for lola to not update the AGENTS.md with transient text (installed skill descriptions for example)
Problem or Use Case
As a project maintainer, I use
lola syncand.lola-reqto recommend AI skill modules to other developers. I also use anAGENTS.mdfile in the repo to provide project specific context and rules to any agent developers are using.When I run
lola sync, it injects the installed skills and some other information into the AGENTS.md. This creates some issues:Category
None
Proposed Solution
I like that and understand why lola injects the information into the agent context. However, I think it would be better if lola wrote that information to a different file, maybe
skills/lola.mdor something, and then updated theAGENTS.mdwith a simple redirect "Check for lola.md and read it if its present"This option could be implemented as a cli arg. I also think it would be nice to have it in a config file (or
.lola-reqalthough im not sure it fits.) because that type of setting would be best applied project wide and not at the user levelAlternatives Considered
No response
Complexity Estimate
I don't know
Additional Context
I really only use
lola sync, so I am unsure if this problem extends to other cli operations