MoonshotAI / MoonshotAI/kimi-code
ignore instructions
Nobody has claimed this yet.
- Dominant language
- TypeScript
- Stars
- 7.5k
- Forks
- 1.2k
- Avg merge
- 11h 53m
- Merged PRs (30d)
- 350
Description
What version of Kimi Code is running?
0.20.1
Which open platform/subscription were you using?
kimi-code
Which model were you using?
kimi-for-coding
What platform is your computer?
Linux 6.19.11-100.fc42.x86_64 x86_64 unknown
What issue are you seeing?
It constantly ignores the instructions in the skills that state that we work together. There states that as the decision maker, I see all the agent's actions and can intervene/help. If the same problem occurs more than # times, it needs to stop, explain the problem, and wait for instructions on how to resolve it. The situation gets even worse after session compression—even agents.md doesn't help.
What steps can reproduce the bug?
Make skill with following instruction, switch to yolo and ask to do some long steps activity with problem in the middle
What is the expected behavior?
No response
Additional information
No response
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Reproduce the behavior in Kimi Code 0.20.1 on Linux by switching to yolo mode and asking for a long activity with a problem in the middle. Compare how skill instructions and agents.md are followed before and after session compression; done means the agent surfaces repeated problems, explains them, and waits for user instructions.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- typescript
- Domain
- ai, cli
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 45/100