session-memory
Scannednpx machina-cli add skill a5c-ai/babysitter/session-memory --openclawSession Memory
Overview
Persistent memory system that survives Claude Code's message compaction. Uses three markdown files in .claude/cc10x/ as a permission-free database for continuity, consistency, and pattern compounding.
Memory Surfaces
- activeContext.md -- Current focus, decisions, learnings, next steps, blockers
- patterns.md -- Project conventions, architecture decisions, common gotchas, reusable solutions
- progress.md -- Task completion tracking with verification evidence
Iron Law
EVERY WORKFLOW MUST:
- LOAD memory at START (and before key decisions)
- UPDATE memory at END (and after learnings/decisions)
Stable Edit Anchors
Safe section headers for Edit operations:
- activeContext:
## Recent Changes,## Learnings,## References - patterns:
## Common Gotchas,## Project SKILL_HINTS - progress:
## Completed,## Verification
Read-Edit-Verify Pattern
- Read file
- Verify anchor exists
- Edit with exact
old_string - Read back to confirm
Tool Rules
- Use
Write()for NEW files (permission-free) - Use
Edit()for EXISTING files (permission-free) - Never use
Write()to overwrite existing files - Never compound commands (
mkdir && cat)
When to Use
- At the start of every CC10X workflow (load)
- At the end of every CC10X workflow (update)
- Before making key decisions (check patterns)
- After discovering learnings or gotchas (persist)
Agents Used
All CC10X agents use this skill. The cc10x-router manages load/update lifecycle.
Source
git clone https://github.com/a5c-ai/babysitter/blob/main/plugins/babysitter/skills/babysit/process/methodologies/cc10x/skills/session-memory/SKILL.mdView on GitHub Overview
Session Memory provides a persistent memory layer across session resets by storing state in three markdown surfaces under .claude/cc10x/. It is designed to support continuity, consistency, and pattern accumulation so CC10X workflows can reference prior decisions and learnings.
How This Skill Works
The skill loads memory at the start of a workflow, reads and validates the three surfaces (activeContext.md, patterns.md, progress.md), and updates them at the end with decisions, learnings, and completion evidence. It relies on a Read-Edit-Verify pattern to ensure anchors exist and edits apply exactly, and uses Write() only for new files and Edit() for existing ones to avoid overwriting content.
When to Use It
- At the start of every CC10X workflow (load memory)
- At the end of every CC10X workflow (update memory)
- Before making key decisions (check patterns and references)
- After discovering learnings or gotchas (persist learnings)
- During workflow design to plan memory surface usage and ensure anchors exist
Quick Start
- Step 1: Ensure activeContext.md, patterns.md, and progress.md exist under .claude/cc10x/
- Step 2: At workflow start, Read each surface and verify anchors exist
- Step 3: At workflow end, Edit the surfaces with new decisions, patterns, and progress; run Verify to confirm
Best Practices
- Always load memory at the very start of a CC10X workflow
- Verify that anchors exist in each surface before editing
- Edit with the exact old_string to avoid unintended changes
- Update memory at the end of the workflow with new learnings and decisions
- Use Write() only for new files and Edit() for existing files
Example Use Cases
- CC10X workflow starts by reading activeContext.md for current focus, decisions, and blockers
- Patterns.md is updated to capture architecture decisions and common gotchas
- Progress.md records completed tasks with verification evidence
- Discovering a new blocker and persisting it to activeContext.md for visibility
- After a run, re-reading all surfaces to verify that anchors and content reflect latest learnings
