v2.3.6 — Fix: Platform-neutral terminology
v2.3.6 — Fix: Platform-neutral terminology
Problem
Workflows, templates, and references used "Claude's Discretion", "Claude's judgment", and generic "Claude" as an agent noun. learnship runs on 6 platforms with different LLMs — this language was wrong on all non-Claude platforms.
Fix
All generic agent language replaced with platform-neutral equivalents:
| Before | After |
|---|---|
| Claude's Discretion | Agent's Discretion |
| Claude's judgment | agent's judgment |
| Claude builds | The agent builds |
| future Claude sessions | future agent sessions |
| consumer is Claude | consumer is the agent |
"Claude Code" (platform name) and "Claude Opus/Sonnet/Haiku" (model names) are unchanged — those are proper nouns.
Affected files
discuss-phase.md, quick.md, templates/context.md, templates/discussion-log.md, references/git-integration.md, references/questioning.md, references/verification-patterns.md + all .windsurf/ copies.
Tests
9 new regression checks (REG-058–066) in validate_regressions.sh §15 enforce platform-neutral terminology going forward · 15/15 suites · 0 failures