/ar:report — write the run up
Turn the record into something a person can read. Do not iterate.
Steps
Read state (cat ./.ar/ar.jsonl | tail -50, plus worklog.md and
results.tsv if present). No ar.jsonl means nothing to report — say so and
stop.
Reconstruct the run per ${CLAUDE_PLUGIN_ROOT}/references/resume-loop.md §1,
across all segments, not just the last: a segment boundary marks a harness
or baseline change, so metrics either side of it are not comparable and the
report must say where the boundaries fell rather than plotting through them.
Pull the winner chain from git — it is an independent record of the same run:
git log --format='%h %s' --grep='"status":"keep"' <config.branch>
git diff <config.snapshotCommit>..<bestCommit> --stat
Write ./.ar/final_report.md:
- Goal and setup — objective, metric, direction, harness, budget.
- Result — baseline → best, absolute and relative, with the confidence
tier. State plainly whether the total gain clears the noise floor; a run
that ended inside its own scatter did not find anything, and the report
says so rather than dressing it up.
- Trajectory — the keep chain, each with commit, metric, delta and the
one-line description of the change.
- What worked — the winning changes and, where the record supports it,
why they worked. Distinguish an explanation from a guess.
- What failed — discarded families of hypotheses. This is the part with
the most reuse value; a future run should not re-derive these.
- Caveats — crashes,
checks_failed iterations, seeds that disagreed,
segment boundaries.
- Next — the strongest untried ideas, from
ideas.md and the plateau
history.
Report the path. .gitignore carries !.ar/final_report.md, so the report is
trackable if it should be committed — offer, do not commit unasked.
Hard rules
- Write
final_report.md and nothing else. No iterating, no commits, no
branch switches, no reverts.
- Every number comes from
ar.jsonl or git. Never recompute or estimate one.
- Report failures and null results as prominently as wins. A loop that banked
noise is a finding; concealing it wastes the next run.
Completion status
End with a terminal status token as the last line of your reply — DONE,
DONE_WITH_CONCERNS, BLOCKED, or NEEDS_CONTEXT — per
references/completion-status.md. For /ar:report, DONE once
final_report.md is written and its path reported; NEEDS_CONTEXT if ar.jsonl
is absent or empty (nothing to summarise).
1---2name: report3description: Write the final summary of an autoresearch run to `./.ar/final_report.md` — trajectory, best commit and its diff, what worked, what failed, and what to try next. Use for `/ar:report`, "write up the autoresearch results", or "summarise the experiment loop". Reads `./.ar/ar.jsonl` from disk, so it works in a fresh session. Writes only the report; never iterates, commits, or reverts, and leaves the loop able to continue.4---56# /ar:report — write the run up78Turn the record into something a person can read. Do not iterate.910## Steps11121. Read state (`cat ./.ar/ar.jsonl | tail -50`, plus `worklog.md` and13 `results.tsv` if present). No `ar.jsonl` means nothing to report — say so and14 stop.152. Reconstruct the run per `${CLAUDE_PLUGIN_ROOT}/references/resume-loop.md` §1,16 across **all** segments, not just the last: a segment boundary marks a harness17 or baseline change, so metrics either side of it are not comparable and the18 report must say where the boundaries fell rather than plotting through them.193. Pull the winner chain from git — it is an independent record of the same run:2021 ```bash22 git log --format='%h %s' --grep='"status":"keep"' <config.branch>23 git diff <config.snapshotCommit>..<bestCommit> --stat24 ```25264. Write `./.ar/final_report.md`:27 - **Goal and setup** — objective, metric, direction, harness, budget.28 - **Result** — baseline → best, absolute and relative, with the confidence29 tier. State plainly whether the total gain clears the noise floor; a run30 that ended inside its own scatter did not find anything, and the report31 says so rather than dressing it up.32 - **Trajectory** — the keep chain, each with commit, metric, delta and the33 one-line description of the change.34 - **What worked** — the winning changes and, where the record supports it,35 why they worked. Distinguish an explanation from a guess.36 - **What failed** — discarded families of hypotheses. This is the part with37 the most reuse value; a future run should not re-derive these.38 - **Caveats** — crashes, `checks_failed` iterations, seeds that disagreed,39 segment boundaries.40 - **Next** — the strongest untried ideas, from `ideas.md` and the plateau41 history.425. Report the path. `.gitignore` carries `!.ar/final_report.md`, so the report is43 trackable if it should be committed — offer, do not commit unasked.4445## Hard rules46471. **Write `final_report.md` and nothing else.** No iterating, no commits, no48 branch switches, no reverts.492. Every number comes from `ar.jsonl` or git. Never recompute or estimate one.503. Report failures and null results as prominently as wins. A loop that banked51 noise is a finding; concealing it wastes the next run.5253## Completion status5455End with a terminal status token as the last line of your reply — `DONE`,56`DONE_WITH_CONCERNS`, `BLOCKED`, or `NEEDS_CONTEXT` — per57`references/completion-status.md`. For `/ar:report`, `DONE` once58`final_report.md` is written and its path reported; `NEEDS_CONTEXT` if `ar.jsonl`59is absent or empty (nothing to summarise).
emaballarin