Feature: Contextual Help - Proactive SlashSense Agent Suggestions

[shakestzd]

Problem

Users don't discover SlashSense agents because:

• No in-context help or suggestions
• Documentation not surfaced at point of need
• Users default to manual workflows or general-purpose agents

Real Example: I manually created worktrees and spawned agents, unaware that SlashSense could orchestrate everything automatically.

Impact

• 🔍 Discovery: Features go unused
• 💰 Cost: Users waste money on expensive alternatives
• ⏱️ Time: Manual workflows are slower
• 📚 Documentation: Excellent docs but not discoverable

Root Cause

SlashSense is reactive (waits for user to ask), not proactive (suggests optimal approach).

Proposed Solution

Add contextual triggers to intent-recognition skill:

Pattern Detection

User Says	SlashSense Suggests
"parallel tasks"	💡 Try /slashsense:parallel-plan for optimized parallel execution
"run tests"	💡 Use test-runner agent (87% cheaper than general-purpose)
"git worktree"	💡 worktree-manager handles this automatically
"create GitHub issue"	💡 issue-orchestrator does bulk operations
"spawn 5 agents"	💡 Show cost comparison (Haiku vs Sonnet)

Example UX

User: "I need to create 5 git worktrees for parallel development"

SlashSense (auto-suggests):
  💡 I can help! SlashSense has a worktree-manager agent that:
  - Creates worktrees automatically
  - Validates configuration
  - Handles cleanup
  - Costs only $0.008 vs $0.06 manual approach

  Want me to use it? (yes/no)

Implementation

1. Trigger Patterns

Add to .claude/skills/intent-recognition.md:

## Contextual Triggers

When user message contains:

**Parallel/Concurrent Work**:
- Patterns: "parallel", "concurrently", "at the same time", "5 tasks"
- Suggestion: "Consider SlashSense parallel workflow for 82% cost savings"
- Command: `/slashsense:parallel-plan`

**Testing**:
- Patterns: "run tests", "pytest", "test suite"
- Suggestion: "Use test-runner agent (87% cheaper)"
- Agent: `slashsense:test-runner`

**Git Worktrees**:
- Patterns: "worktree", "parallel branch", "multiple branches"
- Suggestion: "worktree-manager automates this"
- Agent: `slashsense:worktree-manager`

**GitHub Issues**:
- Patterns: "create issue", "bulk issues", "issue templates"
- Suggestion: "issue-orchestrator handles batch operations"
- Agent: `slashsense:issue-orchestrator`

**Task Delegation**:
- Patterns: "spawn agent", "delegate", "run in parallel"
- Suggestion: "Show cost comparison: Haiku vs Sonnet"
- Skill: `agent-cost-advisor`

2. Suggestion Format

When triggered:
1. Briefly explain SlashSense capability
2. Show cost/time savings
3. Offer to use it
4. Don't be pushy (user can decline)

3. User Preferences

Allow users to:

• Enable/disable suggestions
• Set suggestion frequency (always/sometimes/never)
• Customize trigger keywords

Acceptance Criteria

• Detects common patterns in user messages
• Suggests relevant SlashSense capabilities
• Shows cost/time savings
• Non-intrusive (user can decline)
• Respects user preferences
• Tracks suggestion acceptance rate

Estimated Effort

3-4 hours

Benefits

• Increases SlashSense feature adoption
• Users save money automatically
• Reduces manual workflow friction
• Self-educating system

Example Metrics

After implementation, track:

• Suggestion trigger rate
• User acceptance rate
• Cost savings from accepted suggestions
• Most popular suggested features

Priority: High
Labels: enhancement, ux, discovery