Skip to content

[Auto] session_failure: The request was malformed and could not be processed. This message was blocked b #648

Open
Open
[Auto] session_failure: The request was malformed and could not be processed. This message was blocked b#648
Labels
auto-reportedbugSomething isn't working

Description

@igorcosta
igorcosta
opened on Oct 6, 2026

Auto-Reported Autohand CLI Error

Field Value
Error Type session_failure
CLI Version 0.9.8
Model auto
Provider autohandai
Platform darwin 27.0.0
Session ID af14bfe8-adda-4528-bdb2-11c42c767c6a-1791269785473
Report ID jVBAi7c7cRRKA8eoHtwwq

Error Message

The request was malformed and could not be processed.
This message was blocked by Autohand's safety check. Rephrase it and try again.

Stack Trace

ApiError: The request was malformed and could not be processed.
This message was blocked by Autohand's safety check. Rephrase it and try again.
    at buildFriendlyError (/$bunfs/root/autohand-macos-arm64:87190:26)
    at async makeRequest (/$bunfs/root/autohand-macos-arm64:87109:42)
    at async complete (/$bunfs/root/autohand-macos-arm64:87042:48)
    at processTicksAndRejections (native:7:39)

Session Context

  • Conversation length: 46 messages
  • Context usage: 28%
  • Last tools used: ask_followup_question, ask_followup_question, ask_followup_question
  • Retry attempt: 0 / 3

Additional Context

{
  "clientType": "cli",
  "retryAttempt": 0,
  "maxRetries": 3
}

Auto-reported by Autohand CLI v0.9.8 | Report ID: jVBAi7c7cRRKA8eoHtwwq

Activity

  1. igorcosta
    added
    bugSomething isn't working
    auto-reported
    on Oct 6, 2026

  2. igorcosta commented on Oct 8, 2026

    @igorcosta
    igorcosta
    on Oct 8, 2026
    ContributorAuthor

    Triaged report jVBAi7c7cRRKA8eoHtwwq (CLI 0.9.8, Autohand AI auto). The report records an upstream safety-check rejection, but does not include the request context or inference-side decision needed to establish whether this was a false positive or a client payload problem.

    Keeping this issue open. Next step: correlate this report/session ID with inference logs and reproduce using a sanitized request. The scoped feedback/reporting work improves future diagnostics; it does not establish a fix for this safety-check response.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    auto-reportedbugSomething isn't working

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions