Reflect

sw:reflect — analyze the current session and extract learnings to the ## Skill Memories section of CLAUDE.md. Also exposes a dashboard of skill-refinement suggestions and a session-end nudge (added in increment 0671).

Usage

"Extract learnings from this session"

Subcommands:

sw:reflect                    # Manual reflection — scan session signals and persist learnings
sw:reflect "Focus on the API patterns we discussed"

sw:reflect --on               # Enable auto-reflection via the stop hook
sw:reflect --off              # Disable auto-reflection
sw:reflect --status           # Show config, learning stats, and refinement suggestions
sw:reflect --clear --skill frontend
sw:reflect --clear --all

---

--status dashboard

sw:reflect --status prints the current reflection configuration, a summary of saved Skill Memories, recent learnings, and a Skill Refinement Suggestions section (new in 0671):

REFLECT: Status Dashboard

CONFIGURATION
  Reflection:      Enabled
  Auto-nudge:      Enabled
  Model:           haiku
  Max/session:     3

SKILL MEMORIES (CLAUDE.md)
  Skill           Learnings
  devops          1
  frontend        2
  backend         3
  general         2
  Total:          8

## Skill Refinement Suggestions
  Skill            Signals (J/R/C)   Severity   Last seen         Command
  sw:architect     5 (3/1/1)         high       2026-04-19 18:22  sw:skill-refine sw:architect
  sw:pm            3 (2/1/0)         medium     2026-04-18 09:07  sw:skill-refine sw:pm

Skill Refinement Suggestions

• Skills with ≥3 type: "refinement" signals in .specweave/state/skill-signals.json appear in this section.
• Ranking is severity × recency. Severity weights: high=3, medium=2, low=1.
• Signal sources (J/R/C) are judge-llm / rubric / code-reviewer.
• The entire section is omitted when no skills meet the ≥3 threshold.
• The Command column is copy-pasteable — see sw:skill-refine for full flag and attribution docs.

---

Session-end Nudge

Added in increment 0671. When /sw:done closes an increment, the reflect stop-hook checks whether the session accumulated either:

• ≥1 high-confidence learning (eligible for CLAUDE.md persistence), or
• ≥1 refinement signal attributed to an existing skill.

If so, a one-line prompt is printed at closure:

Detected: 2 signals for sw:architect — run sw:skill-refine sw:architect? (y/N)

or, for a learning:

Detected: 1 high-confidence learning — run sw:reflect? (y/N)

The nudge never auto-executes a command. The user must run the suggested command explicitly. Declining (N or timeout within 5s) is free — the underlying signals remain in skill-signals.json for later.

Disable

// .specweave/config.json
{
  "reflect": {
    "enabled": true,
    "autoNudge": false
  }
}

The close-time check adds under 100ms to /sw:done.

---

Configuration

{
  "reflect": {
    "enabled": true,
    "autoNudge": true,
    "model": "haiku",
    "maxLearningsPerSession": 3
  }
}

Setting	Default	Description
enabled	true	Master switch for reflection
autoNudge	true	Session-end nudge when learnings or refinement signals are detected (0671)
model	"haiku"	LLM model for extraction
maxLearningsPerSession	3	Max learnings persisted per session

---

See also

• sw:skill-refine — the command the nudge and --status dashboard point users toward.
• The full reflect behavior (signal detection rules, quality checklist, what to reject) is documented in plugins/specweave/commands/reflect.md.