Dev-Log - Automatic Development Session Logging
Automatically captures development work from conversation context. One file per project with timestamped entries. No questions asked - just run /dev-log and it extracts everything.
When to Use
- After substantive work (20+ minutes of development)
- When installing/configuring new tools or SDKs
- When creating accounts for services
- When making key architectural decisions
- Before switching projects (capture context while fresh)
- Before ending a session
When NOT to Use
- Reading documentation (no work produced)
- Quick fixes under 10 minutes
- Exploratory sessions with no concrete outcome
Storage Location
Root: ~/dev-log/ (configurable)
First Run Setup
When you first run /dev-log, if the directory doesn't exist, you'll be asked:
"Where would you like to store development logs?"
Options:
- Default location (
~/dev-log/) - Recommended for personal use across all projects - Current project (
./.dev-log/) - Keep logs within project directory - Custom path - Specify your own location
Your choice is saved to .claude/dev-log.local.md in the current project.
Directory Structure
dev-log/
├── projects/
│ ├── clawdbot.md # All clawdbot work
│ ├── dev-log.md # This skill's development
│ └── <project>.md # One file per project
├── drafts/ # Content creation outputs
└── index.json # Project metadata
Quick Reference
| Command | Action |
|---|---|
/dev-log | Auto-capture session to detected project |
/dev-log <project> | Auto-capture to specific project |
/dev-log show | View current project entries |
/dev-log show <project> | View specific project entries |
/dev-log list | List all projects |
/dev-log interview | 12-question content capture (for publishing) |
/dev-log content | Generate content from existing logs |
/dev-log drafts | List content drafts |
/dev-log help | Show detailed help |
Default Mode (Auto-Capture)
Running /dev-log automatically extracts from conversation:
- Summary: What was accomplished
- Decisions Made: Key choices with rationale
- Tools & SDKs: CLIs, libraries, services used
- Accounts & Credentials: Signups, API keys created
- Configuration: Files created/modified
- Key Commands: Significant bash commands run
- Problems Solved: Issues and resolutions
- Next Steps: Pending work identified
- Status: Working/In Progress/Blocked/Exploratory
No questions asked. Just analyzes the conversation and logs.
Interview Mode (Content Creation Only)
Use /dev-log interview ONLY when you want to create publishable content.
Generates:
- Twitter/X thread (7-12 tweets)
- LinkedIn post (~2000 chars)
- Screenshot script
Project Detection
Auto-detects from:
- Git remote URL
- Git repository root
- Current working directory
Override: /dev-log my-project-name
Example Entry
## 2026-01-15 18:26 AEDT
### Summary
Set up Clawdbot with Tailscale Funnel for Gmail webhooks...
### Decisions Made
- **Tailscale over ngrok**: Persistent URL, no signup...
### Tools & SDKs
| Tool | Purpose | Notes |
|------|---------|-------|
| Tailscale | Public HTTPS | Funnel enabled |
### Next Steps
- [ ] Auth ElevenLabs CLI