Render Superconscious ReasoningRun traces in AgentTerm operator console

[mdheller]

Purpose

Superconscious now emits governed recursive reasoning artifacts, and SourceOS reasoning contracts are being promoted into SourceOS-Linux/sourceos-spec.

AgentTerm should render Superconscious run artifacts as an operator-facing trace/evidence console.

Scope

• Render ReasoningRun task state, status, safe trace posture, replay class, and benchmark result.
• Render ReasoningEvent timeline as an operator-readable tree.
• Render coordination records for workspace, grants, policy, model route, skill, tool, memory, approval, and evidence.
• Add a fixture using Superconscious M1 artifacts.
• Keep M1 read-only; no terminal action execution.

Acceptance criteria

• AgentTerm can inspect a Superconscious run directory or exported artifact set.
• The operator can see task, events, evidence refs, replay posture, and benchmark status.
• No raw private reasoning content is required or displayed.
• Docs clarify that AgentTerm is an operator surface, not the cognition loop.

Related repos

• SocioProphet/superconscious
• SourceOS-Linux/sourceos-spec
• SourceOS-Linux/sourceos-devtools
• SocioProphet/agentplane