Define a handful of taxonomy terms

[julianna-ciq]

Summary

• define the accepted taxonomy terms Skill, Autonomy level, AI agent bill of materials (AI-BOM), Kill switch, Reversibility, Handoff, and Context exhaustion
• add the accepted Security & Privacy terms Agent derailment, Agent misuse, Agent sabotage, and Rogue agent
• preserve deferred category and hierarchy fields while adding the agreed aliases, workgroups, related terms, and contrasts
• keep Universal terms unassigned with workgroups: []

Why

These terms were retained through workstream review, and this PR records the presented definitions and scope notes in the taxonomy. The explicit links distinguish adjacent concepts such as Skill and Tool, Kill switch and Reversibility, and the derailment/misuse/sabotage family. Rogue agent is linked to Agent derailment, Agent sabotage, and Agent misuse.

Impact

The taxonomy UI will display concrete definitions for eleven accepted terms, include the four Security & Privacy entries, recognize Hand-off and AI-BOM as aliases, and expose the relevant cross-term relationships.

Validation

• git diff --check
• node --check taxonomy/taxonomy-data.js
• validated the eleven requested entries field by field and confirmed all 25 taxonomy terms remain unique