Issue Station
Sign In
Tools hereToolbox

openclaw - 💡(How to fix) Fix [Bug]: Default CLI template (openclaw.json) is missing required reasoning/minimal flags for 2026.4.2-beta.1 executor [1 comments, 2 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#59936•Fetched 2026-04-08 02:38:36
View on GitHub
Comments
1
Participants
2
Timeline
3
Reactions
0
Author
ashish797ashish797
Participants
ashish797ashish797h3clikejavah3clikejava
Timeline (top)
labeled ×2commented ×1

The default openclaw.json generated by the CLI lacks reasoning_effort and format flags required by the 2026.4.2-beta.1 agent executor, resulting in a silent failure on Telegram with the fallback error message.

Error Message

The default openclaw.json generated by the CLI lacks reasoning_effort and format flags required by the 2026.4.2-beta.1 agent executor, resulting in a silent failure on Telegram with the fallback error message. The agent executor should handle standard instruct-based models (non-reasoning) using the default configuration without throwing a trace-parsing exception. The executor fails to initialize the reasoning-trace parser for the non-reasoning model, returning the Telegram fallback error: Telegram Error Message:

Root Cause

The default openclaw.json generated by the CLI lacks reasoning_effort and format flags required by the 2026.4.2-beta.1 agent executor, resulting in a silent failure on Telegram with the fallback error message.

Code Example

Telegram Error Message:
"Something went wrong while processing your request. Please try again, or use /new to start a fresh session."

Internal Log Trace:
[tools] read failed: ENOENT: .../MEMORY.md (Triggered by the failed reasoning parser state).
RAW_BUFFERClick to expand / collapse

Bug type

Regression (worked before, now fails)

Beta release blocker

Yes

Summary

The default openclaw.json generated by the CLI lacks reasoning_effort and format flags required by the 2026.4.2-beta.1 agent executor, resulting in a silent failure on Telegram with the fallback error message.

Steps to reproduce

Install [email protected] via pnpm.

Run openclaw setup to generate a fresh ~/.openclaw/openclaw.json.

Configure a Telegram channel and set provider keys in ~/.openclaw/.env.

CRITICAL: Set a standard non-reasoning model (specifically tested with openrouter/stepfun/step-3.5-flash:free) as the primary agent model.

Launch the gateway: pnpm openclaw gateway --force.

Send any message to the bot via Telegram.

Expected behavior

The agent executor should handle standard instruct-based models (non-reasoning) using the default configuration without throwing a trace-parsing exception.

Actual behavior

The executor fails to initialize the reasoning-trace parser for the non-reasoning model, returning the Telegram fallback error:

"Something went wrong while processing your request. Please try again, or use /new to start a fresh session."

OpenClaw version

2026.4.2-beta.1 (7cea7c2)

Operating system

Linux (Ubuntu/Debian)

Install method

pnpm / Node.js (v22.22.2)

Model

openrouter/stepfun/step-3.5-flash:free

Provider / routing chain

OpenRouter

Additional provider/model setup details

The issue is highly reproducible with "Flash" or "Lite" models that do not natively output reasoning blocks. The 2026.4.2-beta.1 core engine seems to expect a reasoning-capable schema by default, but the CLI fails to provide the minimal configuration flags needed to bridge this. Confirmed Fix: Injecting these flags manually into agents.defaults.model: JSON "reasoning_effort": "high", "format": "thinking-xml-raw"

Logs, screenshots, and evidence

Telegram Error Message:
"Something went wrong while processing your request. Please try again, or use /new to start a fresh session."

Internal Log Trace:
[tools] read failed: ENOENT: .../MEMORY.md (Triggered by the failed reasoning parser state).

Impact and severity

Affected users/systems: All users using standard/lite models on a fresh install.

Severity: Beta Blocker. Blocks the entire agent workflow for users on free-tier or fast-tier models.

Frequency: Always reproducible with non-reasoning models.

Consequence: Agent-user interaction is completely disabled on chat channels.

Additional information

Testing indicates that the framework now treats "thinking" logic as a mandatory baseline. The CLI template generator must be updated to include these flags to ensure compatibility with non-reasoning providers like Stepfun.

<img width="665" height="818" alt="Image" src="https://github.com/user-attachments/assets/a709cec2-9ba9-4944-948e-aa6ff9da438d" />

extent analysis

TL;DR

Manually injecting the "reasoning_effort" and "format" flags into the agents.defaults.model configuration is likely to resolve the issue.

Guidance

  • Verify that the openclaw.json file generated by the CLI lacks the required reasoning_effort and format flags.
  • Manually add the following configuration to agents.defaults.model:
    • "reasoning_effort": "high"
    • "format": "thinking-xml-raw"
  • Test the setup with a non-reasoning model to ensure the agent executor initializes correctly.
  • Consider updating the CLI template generator to include these flags by default for compatibility with non-reasoning providers.

Example

"agents": {
  "defaults": {
    "model": {
      "reasoning_effort": "high",
      "format": "thinking-xml-raw"
    }
  }
}

Notes

The provided solution is based on the confirmed fix mentioned in the issue. However, it is essential to note that this might be a temporary workaround until the CLI template generator is updated to include the required flags by default.

Recommendation

Apply the workaround by manually injecting the required flags into the configuration, as this is a confirmed fix for the issue. This will ensure compatibility with non-reasoning models until a more permanent solution is implemented.

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…

FAQ

Expected behavior

The agent executor should handle standard instruct-based models (non-reasoning) using the default configuration without throwing a trace-parsing exception.

#indexing error#inference speed#output truncation#response parsing#generation error

Still need to ship something?

×6

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

Back to top recommendations

TRENDING