What does this skill actually do?

It performs deep forensics on stalled or looping agent tasks to identify the exact point of failure and propose a structured fix plan based on product rules.

What are the three invariants this skill preserves?

The skill ensures that: 1) Productive work continues, 2) Only real blockers stop work, and 3) No infinite loops are created by recovery logic.

Does this skill modify my code automatically?

No, this is a diagnostic and planning tool. It produces a written root cause and a proposed plan that requires board or CTO approval before any code is written.

How does it handle cross-company issues?

It respects API boundaries and company scoping; if a token returns a 403, it will request an approved diagnostic path or proceed from available evidence while marking it as provisional.

When should I use this skill?

Use it when an issue tree is stalled, looping, or 'went too deep' and you need to understand the architectural reason why the automation failed to progress.

Diagnose Why Work Stopped | Claude Code Skill

This skill provides a specialized diagnostic framework for investigating why autonomous agents have stopped making progress, stalled, or entered infinite loops within the Paperclip ecosystem. It performs deep forensics on issue trees, identifies specific stop-points, and proposes high-level product rules to ensure productive work continues while preventing recursive failures. Instead of immediate code fixes, it generates comprehensive, multi-phase plans gated by human approval, ensuring that all proposed changes align with system invariants and existing execution semantics.

주요 기능

01Alignment with existing execution semantics documentation

02Deep forensics on stalled issue trees and blocker chains

03Classification of non-progressing issues into actionable categories

04Structured planning with phased subtasks and specialty assignments

05Invariant validation for productive work and loop prevention

0670,206 GitHub stars

사용 사례

01Proposing new product-level liveness rules for agent orchestration

02Investigating why an agent task stalled without a clear blocker

03Resolving infinite loops in automated recovery paths

Tieline — product intent grounded in code

Work Stoppage & Loop Diagnostic

주요 기능

사용 사례

Background Coding Agent

Openfoam Cfd Copilot

Agent Skill Creator

주요 기능

사용 사례