Reflect
Mine the current conversation for durable learnings, then route them into skill edits.
When to invoke
- The user said "reflect" or "/reflect".
- A complex task (5+ tool calls) just landed cleanly and the recipe is worth keeping.
- The agent hit dead ends, found the working path, and the path generalizes.
- The user corrected the agent's approach mid-task.
- A non-trivial workflow emerged that isn't captured anywhere.
Skip when the conversation is trivial, off-topic, or already covered by an existing skill the parent followed correctly. One-offs are not learnings.
Process
1. Locate the active transcript
The parent finds its own transcript file before fanning out. The system prompt names the active workspace's
directory; use that path. Do not glob across
. That crosses workspace boundaries and reads private chats from unrelated projects.
bash
ls -t <agent-transcripts>/*.jsonl <agent-transcripts>/*/*.jsonl <agent-transcripts>/*/subagents/*.jsonl 2>/dev/null | head -10
Three transcript layouts: legacy flat (
), current nested (
), and subagent (
<parent>/subagents/<child>.jsonl
).
For each candidate, read the first JSONL line and check that
contains the conversation's opening user prompt. Take the matching path. If no path resolves, write a tight digest of the session and pass that instead.
2. Spawn three reviewers in parallel
One message, three
calls,
subagent_type: generalPurpose
, explicit
on each, agent mode (
). Reviewers need MCP access for context lookups (tickets, chat threads, observability traces referenced in the transcript); readonly strips MCPs. The prompt forbids file writes; the parent applies edits.
| Lens | | Prompt template |
|---|
| Judgment | your configured reflect-judgment model (default claude-opus-4-8-thinking-xhigh
| references/judgment-reviewer.md
|
| Tooling | your configured reflect-tooling model (default ) | references/tooling-reviewer.md
|
| Divergent | your configured reflect-judgment model (default claude-opus-4-8-thinking-xhigh
| references/divergent-reviewer.md
|
Pass each template verbatim, substituting the transcript path or digest where marked. Reviewers return findings in the
response body.
3. Synthesize
One
call,
subagent_type: generalPurpose
, using your configured reflect-judgment model (default
claude-opus-4-8-thinking-xhigh
), agent mode (
). The synthesizer's quality check includes spot-verifying citations, which can require MCP access; readonly strips MCPs. Use
references/synthesizer.md
verbatim, with each reviewer's full output inlined where marked. The synthesizer returns a structured Accepted / Rejected / Backlog list.
4. Structural enforcement check
Sanity-check the synthesizer's Accepted list. For any item that would be enforced more reliably by a lint rule, script, metadata flag, or runtime check, move it from Accepted to Backlog. The synthesizer already applies this criterion; this is a final pass before edits land. See the encode-lessons-in-structure principle skill.
5. Apply
Before applying any Accepted edit, present the synthesizer's full Accepted/Rejected/Backlog output to the user and wait for explicit approval. The user picks which subset to apply and may redirect routings. Skill changes affect every future agent in the org; do not auto-apply.
Backlog items file to whatever devex / backlog tracker your team uses automatically. Those are tracker submissions, not skill edits. Only the Accepted list waits for approval.
For each approved Accepted item, follow the Routing field exactly:
- Trivial existing-skill edit (a one-line bullet, a tightened sentence, a stale fact corrected): parent does directly.
- Substantive existing-skill edit (a new section, a new pattern table, more than ~10 lines): hand to Cursor's built-in skill and run its draft / test / iterate loop.
tune description: <skill path>
new skill via create-skill: <kebab-name>
If your environment ships a SKILL.md validator, run it on every touched skill before declaring done. Skip this step if it doesn't.
6. Summarize for the user
Short list, no preamble:
- Edits applied: . What changed, one line each.
- New skills created: . One line each (rare).
- Backlog filed to the devex tracker: (). One line each.
- Dropped: one line per rejected finding + reason from the synthesizer.