Memory System Skill

Triggers

Load when the user invokes any of these; otherwise do not load.

Command	Description
/aid-init	Initialize AID system / memory system (first time setup)
/aid-start [role] [phase]	Start a work session
/aid-status	Show current session state
/aid-end	Complete phase gate, collect feedback
/aid-improve	Run improvement analysis (spawns sub-agent)
/aid-memory	Manage memory entries
/aid-reset	Reset memory system (clear state & feedback)

Enums

Roles: product-manager (pm), developer (dev), qa-engineer (qa), tech-lead (lead), data-scientist (ds)

Phases: discovery (Phase 0), prd (Phase 1), tech-spec (Phase 2), breakdown (Phase 3), development (Phase 4), qa-ship (Phase 5)

Sub-Agent: Memory Analysis

On /aid-improve, spawn the memory-analysis-agent:

Task(
  subagent_type: "general-purpose",
  model: "opus",
  prompt: [Read ../../agents/memory-analysis-agent.md],
  description: "Memory system improvement analysis"
)

Pass:

{{FEEDBACK_DATA}} - Contents of ~/.aid/feedback/pending/*.json
{{CURRENT_SKILLS}} - Section headers from skills/memory-system/references/
{{TREND_DATA}} - Contents of ~/.aid/metrics/trends.json
{{MEMORY_ENTRIES}} - Current AID:* Claude Memory entries

Agent returns:

{
  "suggestions": [...],
  "memory_candidates": [...],
  "trends_analysis": {...}
}

Session Start Flow (/aid-start)

1. Check AID directory:

IF ~/.aid/ does not exist:
  → Run /aid-init flow
  → Create directory structure
  → Initialize with defaults
  → Display: "🚀 AID Memory System initialized!"

2. Load state: read ~/.aid/state.json; check pending_feedback_count; check sessions_since_last_improvement.
3. Check improvement suggestion:

IF pending feedback >= threshold:
  Display:
  "📊 I have {pending} feedback items waiting for analysis.
   Would you like to review insights and improve skills?
   [Yes, let's improve] [No, continue working]"

4. Determine role & phase:

IF /aid-start <role> <phase> provided:
  → Use provided values
ELIF state.last_session exists:
  → Suggest: "Continue as {last_role} in {last_phase}?"
ELSE:
  → Ask for role and phase

5. Load skills from skills/memory-system/references/, in order:

1. roles/{role}/SKILL.md
2. roles/{role}/cumulative.md
3. phases/{phase}/SKILL.md
4. phases/{phase}/cumulative.md

6. Update state & begin:

Update ~/.aid/state.json with:
- current_session.active = true
- current_session.role = role
- current_session.phase = phase
- current_session.started_at = now()

Greet user and ask what they're working on.

Session Work Flow

Track revisions (internal counter):

revision_triggers = [
  "fix", "change", "update", "wrong", "missing", "add",
  "incorrect", "error", "should be", "not right", "revise"
]

On each user message containing trigger:
  revision_count += 1
  Note what needed fixing for feedback

Note patterns (internal):

positive_triggers = ["great", "perfect", "exactly", "good", "excellent"]

On positive trigger:
  Note what worked well for feedback

Phase Gate Flow (/aid-end or phase completion detected)

1. Display summary:

📋 Phase Summary: {phase_display_name}

Completed deliverables:
• {deliverable_1}
• {deliverable_2}

Revisions made: {revision_count}

2. Request rating (MANDATORY):

📊 Quality Rating (1-5):

1 ⭐ - Poor, needs significant improvement
2 ⭐⭐ - Below average, missing important elements
3 ⭐⭐⭐ - Acceptable, worked but room for improvement
4 ⭐⭐⭐⭐ - Good, almost perfect
5 ⭐⭐⭐⭐⭐ - Excellent, exactly what I needed

MUST get rating before proceeding

3. Request qualitative feedback (optional): What worked well? What could be improved? Additional notes?
4. Save feedback file:

Save to ~/.aid/feedback/pending/{timestamp}.json:
{
  "context": { "role": "...", "phase": "..." },
  "metrics": { "rating": N, "revisions": N },
  "qualitative": { "what_worked": "...", "what_didnt": "..." }
}

5. Update state:

Update ~/.aid/state.json:
- statistics.total_sessions += 1
- statistics.pending_feedback_count += 1
- statistics.sessions_since_last_improvement += 1
- current_session.active = false

Improvement Flow (/aid-improve)

1. Gather data: all files from ~/.aid/feedback/pending/; current skills from skills/memory-system/references/; trends from ~/.aid/metrics/trends.json.
2. Spawn analysis agent:

Task(
  subagent_type: "general-purpose",
  model: "opus",
  prompt: [../../agents/memory-analysis-agent.md with variables],
  description: "Analyze feedback and suggest improvements"
)

3. Present suggestions with approval buttons: [Approve] [Edit] [Reject] for each skill update; [Promote to Memory] [Skip] for each memory candidate.
4. Apply changes: update approved skill files; add approved memory entries; archive processed feedback to ~/.aid/feedback/processed/.

Anonymization Rules

NEVER include in feedback:

• ❌ Project name or identifier
• ❌ Company name
• ❌ Domain-specific terms
• ❌ Code snippets
• ❌ User names

ALWAYS include:

• ✅ Role (product-manager, developer, etc.)
• ✅ Phase (discovery, prd, etc.)
• ✅ Rating (1-5)
• ✅ Revision count
• ✅ Generic methodology feedback

State Files

File	Purpose
~/.aid/state.json	Current session state
~/.aid/config.yaml	User configuration
~/.aid/feedback/pending/	Unprocessed feedback
~/.aid/feedback/processed/	Archived feedback
~/.aid/metrics/trends.json	Historical trends

Error Handling

Error	Action
Missing ~/.aid/	Run /aid-init automatically
Missing skill files	Use defaults, log warning
Corrupted state.json	Backup and reset
User skips rating	Gently insist, then save as null

Related

• Agent: ../../agents/memory-analysis-agent.md
• References: skills/memory-system/references/
• Commands: ../../commands/aid-*.md