| name | reflect |
| description | Spawn three parallel review subagents over the active transcript, surface learnings, and route each to a concrete edit on an existing skill. Use when the user says reflect. |
| disable-model-invocation | true |
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. Read the
"Transcripts" section of harness/<detected>.md for where they live and in what
form — it differs on all four harnesses, and on Codex and Copilot CLI there is no
per-workspace directory to list. Scope to the active workspace only; globbing
across all workspaces reads private chats from unrelated projects.
Where the harness does give you a per-workspace directory (Cursor, Claude Code),
newest-first by real modification time:
ls -t <transcripts>/*.jsonl <transcripts>/*/*.jsonl <transcripts>/*/subagents/*.jsonl 2>/dev/null | head -10
On Codex, bound by date under ~/.codex/sessions/<YYYY>/<MM>/<DD>/ and match cwd.
On Copilot CLI, query ~/.copilot/session-store.db by cwd for the session ids,
then read each ~/.copilot/session-state/<id>/events.jsonl.
Three transcript layouts: legacy flat (<id>.jsonl), current nested (<id>/<id>.jsonl), and subagent (<parent>/subagents/<child>.jsonl).
For each candidate, read the first JSONL line and check that message.content[0].text 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 subagent calls, the harness's general-purpose agent type, explicit model: on each, agent mode (readonly: false). 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 | model | Prompt template |
|---|
| Judgment | your configured reflect-judgment model (harness default) | references/judgment-reviewer.md |
| Tooling | your configured reflect-tooling model (harness default) | references/tooling-reviewer.md |
| Divergent | your configured reflect-judgment model (harness default) | references/divergent-reviewer.md |
Pass each template verbatim, substituting the transcript path or digest where marked. Reviewers return findings in the Task response body.
3. Synthesize
One subagent call, the harness's general-purpose agent type, using your configured reflect-judgment model (harness default), agent mode (readonly: false). 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 the harness's skill-authoring tool (
harness/built-ins.md) and run its draft / test / iterate loop.
tune description: <skill path> (the skill exists but didn't trigger when it should have): hand to create-skill and run its description-optimization loop.new skill via create-skill: <kebab-name>: hand creation to create-skill. Do not invent the shape ad hoc.
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:
<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.
