diff --git a/CLAUDE.md b/CLAUDE.md
index 2f34849..08e7f97 100644
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -32,7 +32,8 @@ Automatic loads are NOT enough — Read loads required CLAUDE.md context along t
 
 ## Memory Protocol (Cognitive Memory)
 - Skill: `~/.claude/skills/cognitive-memory/` | Data: `~/.claude/memory/`
-- Session start: Load `~/.claude/memory/CORE.md` and `REFLECTION.md`
+- CORE.md auto-loads via MEMORY.md symlinks (no manual loading needed)
+- Session start: Read `~/.claude/memory/REFLECTION.md` for theme context
 - Auto-store on: bug fixes, git commits (mandatory, --episode), architecture decisions, patterns, configs
 - Always tag: project name + technology + category
 - Session end: prompt "Should I store today's learnings?"
diff --git a/skills/cognitive-memory/SKILL.md b/skills/cognitive-memory/SKILL.md
index d9350f4..89cfcfe 100644
--- a/skills/cognitive-memory/SKILL.md
+++ b/skills/cognitive-memory/SKILL.md
@@ -13,7 +13,7 @@ Cognitive Memory provides persistent, human-readable memory storage as markdown
 - Human-readable markdown files with YAML frontmatter
 - Decay scoring to surface relevant memories and let stale ones fade
 - Episodic session logs for chronological context
-- Auto-curated CORE.md loaded into system prompt
+- Auto-curated CORE.md auto-loaded into system prompt via MEMORY.md symlinks
 - Git-tracked for history, rollback, and diff visibility
 - Relations stored directly in frontmatter (self-contained files)
 - Semantic search via Ollama embeddings (nomic-embed-text)
@@ -44,7 +44,7 @@ Cognitive Memory provides persistent, human-readable memory storage as markdown
 - After discovering a reusable pattern
 - After a successful configuration
 - After troubleshooting sessions
-- At session start (recall relevant memories, load CORE.md + REFLECTION.md)
+- At session start (CORE.md auto-loads via symlinks; read REFLECTION.md for theme context)
 - Periodically (reflect to cluster memories, suggest missing tags)
 
 ## Quick Reference
@@ -267,7 +267,7 @@ python client.py search --min-importance 0.3
 
 ## CORE.md
 
-Auto-generated summary of highest-relevance memories (~3K tokens), loaded into the system prompt at session start. Contains links to the most important active memories grouped by type. Regenerated by the `core` CLI command.
+Auto-generated summary of highest-relevance memories (~1K tokens), auto-loaded into the system prompt at every session via MEMORY.md symlinks. Each project's `~/.claude/projects/<project>/memory/MEMORY.md` symlinks to `~/.claude/memory/CORE.md`. Symlinks are refreshed daily by `cognitive-memory-daily.service` (via `claude-memory-symlinks` script). Regenerated by the `core` CLI command.
 
 ## REFLECTION.md
 
@@ -304,7 +304,7 @@ This skill should be used proactively when:
 4. **Pattern discovered** - Store for future recall
 5. **Configuration worked** - Store what worked and why
 6. **Troubleshooting complete** - Store what was tried, what worked
-7. **Session start** - Load CORE.md + REFLECTION.md, then recall relevant memories (use `--semantic` for deeper matching)
+7. **Session start** - CORE.md auto-loads via MEMORY.md symlinks; read REFLECTION.md for theme context, then recall relevant memories (use `--semantic` for deeper matching)
 8. **Multi-step workflow documented** - Use `procedure` type with structured steps/preconditions/postconditions
 9. **Periodically** - Run `reflect` to cluster memories and surface cross-cutting insights. Run `tags suggest` to find missing tag connections
 10. **After adding memories** - Run `embed` to refresh semantic search index