Claude-skill-registry deciduous
Plan, implement, track, and reflect on your work goals and decisions.
install
source · Clone the upstream repo
git clone https://github.com/majiayu000/claude-skill-registry
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/majiayu000/claude-skill-registry "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/data/deciduous" ~/.claude/skills/majiayu000-claude-skill-registry-deciduous && rm -rf "$T"
manifest:
skills/data/deciduous/SKILL.mdsource content
Planning & Decision Graph Logging
Track every goal, decision, and outcome in the decision graph. This creates persistent memory that survives context loss.
- ALWAYS LOG BEFORE YOU CODE, NOT AFTER.
- Log at the granularity of TODOs or task items.
- When drafting a plan create the GOAL node.
- User Decisions should be tracked
When to Log (Automatic Triggers)
| Situation | Node Type | Example |
|---|---|---|
| In plan mode | | "Add user authentication" |
| TODO / Task Item | | "Implementing JWT auth middleware" |
| User requests new feature | | "Add user authentication" |
| Choosing between approaches | | "Choose between JWT vs sessions" |
| Considering an option | | "Use JWT with refresh tokens" |
| About to write/edit code | | "Implementing JWT auth middleware" |
| Work completed or failed | | "JWT auth working" or "JWT approach failed" |
| Important observation | | "Existing code uses cookie-based sessions" |
Commands
# Create nodes (always include confidence -c) deciduous add goal "Title" -c 90 -p "User's exact request" deciduous add decision "Title" -c 75 deciduous add action "Title" -c 85 deciduous add outcome "Title" -c 90 deciduous add observation "Title" -c 80 # CRITICAL: Link nodes immediately after creation deciduous link <parent_id> <child_id> -r "Reason for connection" # After git commits, link to the graph deciduous add action "Committed feature X" -c 90 --commit HEAD # View the graph deciduous nodes deciduous edges
Rules
- Log BEFORE acting - Create the action node before writing code
- Link IMMEDIATELY - Every node except root goals must have a parent
- Capture verbatim prompts - Use
with the user's exact words for goals-p - Include confidence - Always use
flag (0-100)-c - Log outcomes - Both successes AND failures get logged
Confidence Guidelines
- 90-100: Certain, verified, tested
- 75-89: High confidence, likely correct
- 50-74: Moderate confidence, some uncertainty
- Below 50: Experimental, speculative
The Memory Loop
User Request → Log goal with -p ↓ Choose Approach → Log decision + options ↓ Start Coding → Log action FIRST ↓ Complete Work → Log outcome, link to parent ↓ Git Commit → Log with --commit HEAD
Remember: The decision graph is your persistent memory. Log as you work, not after.