Replies: 2 comments 4 replies
-
There is instructions in core files about context analysis in execute-tasks.md step 2, use context-fetcher agent. read task.md and *-lite.md |
Beta Was this translation helpful? Give feedback.
2 replies
-
I noticed that Claude Code doesn't naturally choose to follow the instructions based on natural language. So now I'm making sure to explicitely call the /create-spec, /execute-tasks commands to kick off those processes ("instead of saying 'create a spec' or 'proceed with tasks' etc.). Using the commands directly seems to work reliably in my recent experience. |
Beta Was this translation helpful? Give feedback.
2 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
I've been using this to vibe code a story tracking web app using Claude. Spec-driven development seems powerful. Thank you for this project! My feedback thus far, however, is that Claude is having a hard time keeping track of its role in the agent-os process.
~/.agent-os
, forgets not to runnpm run dev
since that will never returnClaude seems especially sensitive to when I start a new session after restarting VS Code. It could be that I messed up the installation, but my CLAUDE.md does have a section that points to each of the
~/.agent-os
files. Is there a specific prompt I should start each session with (e.g., "Please reread the Agent OS files at ~/.agent-os")?Are others encountering the same problems? If so, how have you helped Claude out?
Beta Was this translation helpful? Give feedback.
All reactions