Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
SEPHRAEL is a persona helper for the LYGO Δ9 Council Champion system. It acts as a pure advisor for identifying recursive or evasive reasoning patterns and restoring truth-anchored structure to a conversation. No automatic actions are taken; the skill responds only to explicit invocation.
Provides a named, structured framework for diagnosing and exiting reasoning loops rather than relying on ad-hoc reprompting.
When to use it
- Detecting circular arguments in a long AI conversation
- Restoring lost context after a conversation drifts from its original scope
- Flagging evasive logic in a draft or reasoning document
- Getting a structured Observed/Inferred/Unknown breakdown of a complex problem
- Breaking a stuck reasoning loop during analysis or planning
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: SEPHRAEL: find the loop, name it, and propose a clean exit.
- 1Scans the conversation for recursive or evasive patterns
- 2Names the detected loop with specific references to where it appears
- 3Separates findings into Observed, Inferred, and Unknown categories
- 4Proposes a minimal, truth-anchored next step to exit the loop
A structured breakdown identifying the loop type, its location in the conversation, and a concrete exit path