dcdpr / dcdpr/jp

Fable refusals if interrupted / anthropic_thinking_signature not provided

Open
#829 8 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Rust
Stars
16
Forks
3
Avg merge
1d 1h
Merged PRs (30d)
121

Description

The model declined this request (reasoning_extraction): This request was blocked as it seems to violate Anthropic's Terms of Service restrictions on reverse engineering or duplicating model outputs.

Model is anthropic/claude-fable-5.
Reasoning style is "full".
Interrupt strategy is "stop".
User action: Interrupted mid-reasoning.
Consequence: all further prompts fail.

Workaround: never append a turn after interruption, instead undo last turn using jp query --replay --edit and pull a "Minority Report" by appending preemptive correction onto the last turn, so that you can let it complete and get a proper signature.

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Reproduce the failure with anthropic/claude-fable-5 using full reasoning, stop interruption, and a follow-up prompt. Inspect the jp query --replay --edit workflow and the interruption path; done means an interrupted turn no longer causes all subsequent prompts to fail or requires the stated workaround.

Written by the indexing model from the issue text.

Assessment

Tech stack
rust
Domain
ai, cli
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Active
Clarity
Needs clarification
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.