Reflect

Other keel skills named here sit beside this one at ${CLAUDE_SKILL_DIR}/../<name>/SKILL.md, and a principle at principle-<name>. The Skill tool refuses them, so read the file and follow it.

Mine the current conversation for durable learnings, then route them into skill edits.

Invoke when the user says "reflect" or "/keel:reflect". Skip when the conversation is trivial, off-topic, or already covered by an existing skill the parent followed correctly. One-offs are not learnings.

1. Locate the active transcript

Before fanning out, find this conversation's transcript. Claude Code keeps this project's transcripts in ${CLAUDE_CONFIG_DIR:-$HOME/.claude}/projects/<cwd-slug>/, where <cwd-slug> is the session's starting directory with every character other than a letter or digit turned into - (so /Users/you/proj becomes -Users-you-proj). Use that directory only. Do not glob across ${CLAUDE_CONFIG_DIR:-$HOME/.claude}/projects/*/. That crosses project boundaries and reads private chats from unrelated projects.

ls -t ${CLAUDE_CONFIG_DIR:-$HOME/.claude}/projects/<cwd-slug>/${CLAUDE_SESSION_ID}.jsonl ${CLAUDE_CONFIG_DIR:-$HOME/.claude}/projects/<cwd-slug>/*.jsonl 2>/dev/null | head -10

That covers the session (<session-id>.jsonl, one event per line) and its subagents (<session-id>/subagents/agent-<id>.jsonl). The current session's file is ${CLAUDE_SESSION_ID}.jsonl. If that variable shows unexpanded, take the newest candidates. For each candidate, find the first event with "type":"user". Take the file whose message.content contains the conversation's opening user prompt. If none matches, pass a tight digest of the session instead.

2. Spawn three reviewers in parallel

One message, three Agent calls, subagent_type: general-purpose. Reviewers need MCP access to look up the tickets, chat threads, and observability traces the transcript references. Claude Code subagents inherit the parent's MCP tools, so they have it.

Each reviewer and the synthesizer name a role line in ${CLAUDE_CONFIG_DIR:-$HOME/.claude}/keel-models.md and a default. Read ${CLAUDE_CONFIG_DIR:-$HOME/.claude}/keel-models.md if it exists; use the role's line, else the default. Pass that value as the Agent call's model. Omit model when the value is inherit, inherit-parent, or auto. If the Agent tool rejects a value, use the default and say so. If it rejects the default, omit model and say so. Pick each subagent's effort for the job, per the Subagents section of the lead skill.

Pass each template verbatim, substituting the transcript path or digest where marked. Reviewers return findings in their final report.

3. Synthesize

One Agent call, subagent_type: general-purpose, model from the reflect judgment, divergent, synthesizer line (default opus). It spot-verifies citations through the MCP tools it inherits from the parent. Pass references/synthesizer.md verbatim, with each reviewer's full output inlined where marked. It returns an Accepted / Rejected / Backlog list.

4. Structural enforcement check

Move any Accepted item that a lint rule, script, metadata flag, or runtime check would enforce more reliably to Backlog. See the encode-lessons-in-structure principle skill.

5. Apply

Present the synthesizer's full Accepted / Rejected / Backlog output and wait for explicit approval before applying any Accepted edit. The user picks the subset and may redirect routings. Skill changes affect every future agent in the org. Do not auto-apply.

File each Backlog item to your team's devex or backlog tracker without waiting. Only the Accepted list waits for approval.

Never edit a skill under plugins/cache/. Before approval, show its destination, .claude/skills/<skill>/ or ${CLAUDE_CONFIG_DIR:-$HOME/.claude}/skills/<skill>/. After approval, copy the skill folder there and edit the copy. Never open a PR or push.

Follow each approved row's Routing exactly:

• Trivial existing-skill edit (a one-line bullet, a tightened sentence, a stale fact corrected): the parent does it directly.
• Substantive existing-skill edit (a new section, a new pattern table, more than ~10 lines): hand to the anthropic-skills:skill-creator skill and run its draft / test / iterate loop. Without it, follow the lead skill's Authoring a skill playbook (${CLAUDE_SKILL_DIR}/../lead/playbooks/authoring-a-skill.md).
• tune description: <skill path> (the skill exists but didn't trigger when it should have): hand to skill-creator and run its description-optimization loop.
• new skill via skill-creator: <kebab-name>: hand creation to skill-creator. Do not invent the shape ad hoc.

If your environment ships a SKILL.md validator, run it on every touched skill before declaring done. Inside a Claude Code plugin, claude plugin validate <plugin-dir> is one.

6. Summarize for the user

Short list, no preamble:

• Edits applied: <skill path>. What changed, one line each.
• New skills created: <skill path>. One line each (rare).
• Backlog filed to the devex tracker: <issue title> (<tags>). One line each.
• Dropped: one line per rejected finding + reason from the synthesizer.