Reflect

Reflect — Human AI-Usage Self-Assessment

This skill is for you (the human) to reflect on how well you're leveraging AI as a thought amplifier — not a review of agent performance.

Data source: The reflect hook (session-end) automatically collects observations, analyzes patterns, and updates metrics.json. This skill consumes that hook-produced data to produce a human-readable self-assessment.

Hook (auto)                    Skill (on-demand /reflect)
─────────────                  ──────────────────────────
observe → obs/*.jsonl  ──→     epic reflect --context 30
evolve → metrics.json  ──→     5-dimension scorecard
seed → evolved skills  ──→     Action items for the human
ingest → memory graph  ──→     Trend analysis

Iron Law

No score without evidence. Every rating must directly cite at least one of: obs stats, evolution patterns, memory nodes, or session summaries. Block self-serving bias: "doing well" conclusions require concrete metrics.

Process

Step 0 — Collect Context

# Uses Rust subcommand — works on all platforms (Linux, macOS, Windows)
epic reflect --context 30 > /tmp/reflect_ctx.json

Fallback if subcommand fails:

echo "obs_files: $(ls "$HARNESS_DIR/obs/" | wc -l)"
python3 -c "import json; m=json.load(open('$HARNESS_DIR/metrics.json')); print('total_sessions:', m.get('total_sessions',0))"

Query memory (if active):

epic mem recall "AI usage patterns decisions metacognition" --limit 8
epic mem list --type decision --limit 5
epic mem list --type pattern --limit 5

Step 1 — 5-Dimension Reflection

Score each dimension independently: 1–10 + evidence citation + one-line diagnosis.

Dimension 1: Thought Amplification

Question: Is AI a mere executor (code typist) or a genuine thought partner?

Metrics:

• Agent tool call ratio (Agent / total_obs — higher = delegated thinking)
• Skill invocation frequency (meta-layer usage)
• council/discover/spec execution history
• Memory decision node count

Scoring:

Dimension 2: Self-Improvement

Question: Learning from mistakes, or repeating the same patterns?

Metrics:

• evolution_stats.pattern_frequency — recurring patterns?
• evolution_stats.stagnation_count — stagnant sessions
• evolution_stats.trend_last10 — improving/stable/declining distribution
• Evolved skills count vs timespan

Scoring:

Dimension 3: Metacognitive Expansion

Question: Are conversations with AI helping recognize and upgrade your own thinking?

Metrics:

• Memory concept/pattern node count and recency
• Decision/ADR recording frequency
• Session snapshots mentioning "learnings"
• /discover /spec execution history (problem-framing practice)

Scoring:

Dimension 4: Prompt Engineering

Question: Is prompt quality improving over time?

Metrics:

• output_quality dimension average trend (metrics score_history)
• tool_success rate trend
• Evolved skills with prompt improvement patterns
• Session average score trajectory (early vs recent)

Scoring:

Dimension 5: Execution Efficiency

Question: Achieving the same goals faster and cheaper through AI?

Metrics:

• execution_cost dimension average (1.0 = optimal)
• Bash dominance (Bash > 50% suggests excessive low-level repetition)
• Context compaction frequency (too frequent = context waste)
• Agent sub-agent parallel usage

Scoring:

Step 2 — Summary Scorecard

## AI Thought-Amplifier Reflection Report
Generated: {ISO-8601}  |  Window: {N} days  |  Total sessions: {total_sessions}

| Dimension            | Score | Grade | Key Evidence |
|----------------------|-------|-------|--------------|
| Thought Amplification | X/10 | 🔴/🟡/🟢 | Agent {N}x ({P}%), Skill {M}x |
| Self-Improvement      | X/10 | 🔴/🟡/🟢 | trend={T}, stagnation={S}x |
| Metacognitive Expansion | X/10 | 🔴/🟡/🟢 | decisions {D}, session notes {M} |
| Prompt Engineering    | X/10 | 🔴/🟡/🟢 | output_quality={Q}, trend={Δ} |
| Execution Efficiency  | X/10 | 🔴/🟡/🟢 | execution_cost={C}, Bash%={B} |
| **Overall**           | **X/10** | | |

Grade: 🟢 8–10 (good)  🟡 5–7 (fair)  🔴 1–4 (needs improvement)

Step 3 — Cold Summary

3–5 sentences. Rules:

• Open with the lowest-scoring dimension
• No "doing well" claims without numeric evidence
• State the single biggest bottleneck
• Classify usage as "execution automation" or "thought amplification"

Step 4 — Action Items (minimum 3)

Format: [Priority] Title — Concrete action — Expected impact

### Next Reflection Actions

1. [HIGH] {title}
   - Action: {concrete steps}
   - Metric: {how to measure}
   - Deadline: {session count or date}

2. [MED] ...
3. [LOW] ...

Recommended action pool (select by low-scoring dimensions):

• Thought Amplification low → council mode weekly, /spec writing habit
• Self-Improvement low → /evolve history periodic review, manual pattern notes
• Metacognition low → epic mem add --type decision after every important decision
• Prompt low → post low-quality session prompt review → seed improved evolved skill
• Efficiency low → Agent parallel sub-agent patterns, script repetitive Bash calls

Step 5 — Save to Memory

Save reflection to memory (if mem tools active):

epic mem add \
  --type session \
  --title "AI usage reflection {date}" \
  --tags "reflection,metacognition" \
  --importance 0.8 \
  --body "Overall: {score}/10. Lowest: {lowest_dim}. Top action: {top_action}"

Anti-Rationalization

Evidence Required

• /tmp/reflect_ctx.json or manually collected data exists
• Each of 5 dimensions cites ≥ 1 numeric evidence
• Summary scorecard output complete
• Cold summary states 1 specific bottleneck
• Minimum 3 action items, each with measurement metric
• Anti-Rationalization table applied to low-scoring dimensions (1–4)

Red Flags

• All dimensions ≥ 7 → suspect positive bias. Re-verify each score against evidence.
• Action items at "use more often" level → rewrite with concrete actions and metrics.
• Summary under 200 chars → insufficient analysis. Cite more evidence.
• Script failure stops reflection → fall back to manual collection and continue.
• Memory save omitted → reflection won't carry into next session.

Source: hashgraph-online/awesome-codex-plugins → plugins/epicsagas/epic-harness/skills/reflect/SKILL.md