Issue Station
Sign In
Tools hereToolbox

openclaw - ✅(Solved) Fix Bug: kimi model tool calls missing required parameters (~85% failure rate) [1 pull requests, 11 comments, 3 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
openclaw/openclaw#58817•Fetched 2026-04-08 02:32:24
View on GitHub
Comments
11
Participants
3
Timeline
16
Reactions
0
Author
kennyxkkennyxk
Participants
kennyxkkennyxkpoeppemieke7-delpoeppemieke7-delyuanaichiyuanaichi
Timeline (top)
commented ×10cross-referenced ×3closed ×1locked ×1

Error Message

Validation failed for tool "exec":
  - command: must have required property 'command'
  Received arguments: {}

Validation failed for tool "write":
  - content: must have required property 'content'
  Received arguments: {}

Root Cause

Root Cause (Suspected)

Fix Action

Workaround

Switch to MiniMax or other stable models as default.

PR fix notes

PR #59440: fix: normalize Kimi anthropic tool payloads

Description (problem / solution / changelog)

Summary

  • mark Kimi and Kimi Coding as OpenAI-function-style Anthropic tool payload providers
  • align the bundled plugin capability metadata with the fallback capability registry
  • add regression coverage for Kimi payload normalization on the Anthropic Messages path

Fixes #59327 Fixes #54442 Fixes #58817 Fixes #53747 Fixes #56916

Changed files

  • CHANGELOG.md (modified, +1/-0)
  • extensions/kimi-coding/index.ts (modified, +2/-0)
  • src/agents/pi-embedded-runner-extraparams.test.ts (modified, +71/-8)
  • src/agents/provider-capabilities.test.ts (modified, +13/-6)
  • src/agents/provider-capabilities.ts (modified, +2/-0)

Code Example

Validation failed for tool "exec":
  - command: must have required property 'command'
  Received arguments: {}

Validation failed for tool "write":
  - content: must have required property 'content'
  Received arguments: {}

---

OpenClaw: 2026.3.31
Node: v22.22.0
OS: macOS 26.4 (arm64)
Gateway: running
LCM: enabled=true, threshold=0.75

---

openclaw config set lcm.threshold 0.85
   # Error: Config validation failed: <root>: Unrecognized key: "lcm"
RAW_BUFFERClick to expand / collapse

Bug Description

When using kimi/kimi-code as the default model, tool calls frequently fail with "missing required parameter" errors. The model generates tool call requests without including the required parameters.

Error Message

Validation failed for tool "exec":
  - command: must have required property 'command'
  Received arguments: {}

Validation failed for tool "write":
  - content: must have required property 'content'
  Received arguments: {}

Reproduction

  1. Use kimi/kimi-code as default model
  2. Make 5-10 consecutive tool calls (exec or write)
  3. Observe first batch of calls fail with missing parameters

Impact

ToolFailure Rate
exec~85%
write~90%

Environment

OpenClaw: 2026.3.31
Node: v22.22.0
OS: macOS 26.4 (arm64)
Gateway: running
LCM: enabled=true, threshold=0.75

Comparison

ModelSuccess Rate
kimi/kimi-code~15% ❌
minimax-cn/MiniMax-M2.7100% ✅

Tested Solutions

  1. LCM threshold adjustment - Failed (config key not found)

    openclaw config set lcm.threshold 0.85
# Error: Config validation failed: <root>: Unrecognized key: "lcm"

  2. OpenClaw upgrade - Issue persists after 3.24 → 3.31

  3. Model switch - MiniMax-M2.7 shows 100% success rate

Root Cause (Suspected)

  • LCM (Long Context Memory) compression may be truncating tool call parameters
  • Or kimi model's output is being cut off during rapid consecutive calls

Workaround

Switch to MiniMax or other stable models as default.

Suggested Fixes

  1. Make LCM threshold configurable
  2. Add parameter validation + auto-retry before tool call
  3. Investigate kimi model's tool call parameter generation

Reported by: Kenny (via OpenClaw)

extent analysis

TL;DR

Switch to a stable model like MiniMax-M2.7 as a temporary workaround to avoid "missing required parameter" errors with the kimi/kimi-code model.

Guidance

  • Verify that the issue persists when making consecutive tool calls with the kimi/kimi-code model, and confirm that switching to MiniMax-M2.7 resolves the issue.
  • Investigate the LCM compression settings and their potential impact on tool call parameters, considering the suspected root cause of parameter truncation.
  • Consider implementing parameter validation and auto-retry mechanisms for tool calls to mitigate the issue, as suggested in the issue report.
  • Review the kimi model's tool call parameter generation to identify potential issues or limitations.

Example

No code snippet is provided, as the issue does not include specific code examples that can be used to illustrate a solution.

Notes

The provided information suggests that the issue is specific to the kimi/kimi-code model and may be related to LCM compression or the model's output generation. However, the exact cause is still uncertain, and further investigation is needed to determine the root cause and develop a permanent fix.

Recommendation

Apply the workaround by switching to a stable model like MiniMax-M2.7, as it has been shown to have a 100% success rate in the comparison tests. This will allow for temporary mitigation of the issue while further investigation and development of a permanent fix are underway.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…
#api#autograd error#model save/load#optimization#mixed precision

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING