ai-dlc-blockers

# AI-DLC Blockers

Blockers are obstacles that prevent progress on a task. In AI-DLC, properly documenting blockers is essential for productive iteration—it ensures the next session can address issues directly rather than rediscovering them.

## Why Document Blockers

### The Context Reset Problem

When you run `/clear`:
- Context window is wiped
- Previous attempts are forgotten
- Discoveries are lost

Without blocker documentation:
- Next iteration repeats the same failed attempts
- Same dead ends are explored again
- No learning occurs across iterations

With blocker documentation:
- Next iteration knows what didn't work
- Alternative approaches can be tried
- Progress accumulates across sessions

### The Knowledge Transfer

Blockers are a form of knowledge transfer:
- From current session → next session
- From current hat → next hat (if `/fail` called)
- From AI → human (if intervention needed)

## Blocker Anatomy

### Good Blocker Documentation

```markdown
## Blocker: OAuth Callback Not Working

### What I Tried
1. Set callback URL to `http://localhost:3000/auth/callback`
2. Verified Google Cloud Console configuration
3. Checked for CORS issues - none found

### Where I Got Stuck
The callback is received, but the authorization code exchange fails with:
```
Error: invalid_grant - Code was already redeemed
```

### What I Think Is Happening
The code might be getting used twice - possibly a React strict mode double-render issue?

### Suggested Next Steps
1. Disable React strict mode temporarily to test
2. Add request deduplication on the callback endpoint
3. Check if there's a race condition in the OAuth flow
```

### Bad Blocker Documentation

```markdown
OAuth doesn't work. Tried everything.
```

## Blocker Categories

### 1. Technical Blockers

Issues with code, APIs, or infrastructure:

```markdown
## Blocker: Database Connection Timeout

### Symptoms
- Connection fails after 30 seconds
- Error: ETIMEDOUT

### Attempted Solutions
- Increased timeout to 60s -