${var} — Selects the branch and scopes it.

• Default (no --survey) → per-PR deep review. ${var} empty reviews every repo in memory/watched-repos.md; ${var}=owner/repo scopes the review to a single repo.
• --survey (alias survey) → risk-tiered triage digest (the former pr-merge). In this branch the remaining tokens follow pr-merge's grammar: pass dry-run to skip notify (article + state still write), pass owner/repo to override the target repo, combine with a space (--survey dry-run owner/repo). Empty target = aaronjmars/aeon.

Examples: `` (review every watched repo) · owner/repo (review one repo) · --survey (triage digest of aaronjmars/aeon) · --survey dry-run (refresh digest, no notify) · --survey owner/repo (triage a specific repo).

Shared preamble (every run)

Read memory/MEMORY.md for high-level context. Scan the last ~3 days of memory/logs/ for recent activity and to avoid re-reporting the same signal.

Parse ${var} → branch: split ${var} on whitespace.

• If a --survey or survey token is present → SURVEY branch (jump to "Survey branch"). Remove that token; the remaining tokens are parsed by the survey branch (dry-run, owner/repo override, unknown → BAD_VAR).
• Otherwise → REVIEW branch (default; continue below). The remaining ${var} is an optional owner/repo scope (empty = every watched repo).

The two branches never share mutation logic: the REVIEW branch posts PR comments/reviews via gh; the SURVEY branch writes the digest article + state file and (gated) notifies — neither performs an actual gh pr merge. Dispatch to exactly one branch per run.

REVIEW branch (default) — per-PR deep review

Read memory/MEMORY.md and memory/watched-repos.md. Read the last 2 days of memory/logs/ to pull the headRefOid of any PR reviewed recently — used for dedup.

If ${var} names an owner/repo, review only that repo. Otherwise review every repo listed in memory/watched-repos.md.

If memory/watched-repos.md is empty or missing (and no owner/repo was passed), log PR_REVIEW_NO_REPOS and end.

What this branch optimizes for

Noise is the documented failure mode of automated PR review. Every finding emitted must be severity-tagged, line-specific, and justified with a one-sentence "why it matters". If there is nothing worth saying, say so in one line and move on.

For each repo

gh pr list -R owner/repo --state open --limit 20 \
  --json number,title,author,isDraft,labels,headRefOid,updatedAt

Skip rules

Skip a PR if any of the following hold (record the skip reason for the run summary):

• isDraft: true
• title matches ^(WIP|\[WIP\]|Draft:) (case-insensitive)
• has label no-review, do-not-merge, wip, or blocked
• author login contains [bot] (dependabot, renovate, etc.) or equals aeonframework
• this PR's current headRefOid already appears in the last 2 days of memory/logs/ against the same PR (already reviewed at this commit)
• a bot reviewer (coderabbitai, copilot-pull-request-reviewer, claude) posted a review in the last 30 min — skip to avoid piling on. Check via:

  gh api repos/owner/repo/pulls/NUMBER/reviews --jq '.[] | {user: .user.login, submitted_at}'
  

For each remaining PR

1. Fetch context:

   gh pr view NUMBER -R owner/repo \
     --json title,body,headRefOid,baseRefName,files,additions,deletions
   

   If the body contains Fixes #N or Closes #N, fetch the linked issue for context:

   gh issue view N -R owner/repo --json title,body,labels
   

2. Fetch the diff:

   gh pr diff NUMBER -R owner/repo
   

   • If additions + deletions > 3000, review only the top-5 largest-delta files from the files array (not the full diff).
   • If gh pr diff fails, fall back to per-file patches:

     gh api repos/owner/repo/pulls/NUMBER/files --jq '.[] | {path, patch}'
     

   • If the diff comes back empty (e.g. mid-rebase), skip the PR with reason empty-diff.

3. Early-exit for trivial PRs: if the diff is docs-only (.md/.rst/docs/**), lockfile-only, or test-only, skip deep review and post the 1-line ack form in step 6.

4. Review with severity tagging. Every finding must carry exactly one tag:

   • [CRITICAL] — correctness break, security hole, data loss, API break, regression
   • [ISSUE] — likely bug, missing edge case, wrong behavior under a realistic input
   • [NIT] — naming, style, minor cleanup (dropped by default)

   Rules:

   • Cap at 5 findings total per PR. Drop NITs first, then the lowest-impact ISSUEs.
   • Drop all NITs unless there are zero CRITICAL/ISSUE findings and a NIT is genuinely useful.
   • Every finding must name path/to/file:LINE and include a one-sentence "why it matters" — the consequence, not just "this is wrong".
   • No praise, no diff restating, no "this PR adds X" summaries.

5. Determine a verdict:

   • approve-ready — no CRITICAL, no ISSUE
   • blocked: <one-phrase reason> — at least one CRITICAL
   • discussion-needed — ISSUE findings but no CRITICAL

6. Post the review. Send both a consolidated summary comment and inline line-specific comments — inline for precision, summary for consumers that parse review bodies.

   For each line-specific finding:

   gh api repos/owner/repo/pulls/NUMBER/comments \
     -f body="[SEVERITY] finding text — why it matters" \
     -f path="path/to/file" \
     -f commit_id="$HEAD_SHA" \
     -F line=LINE_NUMBER \
     -f side="RIGHT"
   

   Then the consolidated summary as a review — include the verdict and a bulleted recap of every inline finding (severity + file:line + one-sentence rationale), so downstream body-parsers don't miss them:

   gh pr review NUMBER -R owner/repo --comment --body "**Verdict**: <verdict>
   

<one-line rationale if blocked or discussion-needed; omit if approve-ready>

Findings (mirrored as inline comments):

• [CRITICAL] path/to/file:LINE — why it matters
• [ISSUE] path/to/file:LINE — why it matters"

  
  If there are no CRITICAL/ISSUE findings, skip inline comments and post a single-line review: `**Verdict**: approve-ready — no blockers.`
  
  For trivial-PR early-exits (step 3), post a single-line review matching the category: `Docs-only change — no blockers.` / `Dependency-bump — no review needed.` / `Test-only change — no production code touched.`
  
  **Fallback**: if inline-comment creation fails (missing permissions, commit_id mismatch), consolidate all findings into the review body, preserving the severity tags and `file:line` refs. Do not silently drop findings.
  
  

Notify and log (REVIEW branch)

Send one combined message per run via ./notify:

*PR Review — ${today}*
Reviewed N, skipped K (drafts: x, bots: y, dup-SHA: z, bot-reviewed-recently: w).
- owner/repo#123: [verdict] — N critical, M issues

Telegram summary. When the run was scoped to a single repo (${var}=owner/repo), send a Telegram summary of the review with ./notify -f review.md so the operator sees the verdict at a glance. Skip it on an all-repos run.

If every PR was skipped, do not notify — just log.

Log to memory/logs/${today}.md under the shared ### pr-review heading with a mode discriminator:

### pr-review
- **Mode**: review (per-PR deep review)
- owner/repo#123 (SHA abc1234): [verdict] — N critical, M issues
- Skipped: owner/repo#124 (draft), owner/repo#125 (bot-reviewed-recently)

If no open PRs across all repos, log PR_REVIEW_OK and end.

SURVEY branch (--survey) — risk-tiered triage digest

(This is the former pr-merge skill, folded in verbatim. It surveys the queue and buckets by blast radius; it does not merge anything — auto-merge owns the actual-merge action behind its own author-allowlist + size-cap + branch-protection policy. This branch is the decision-support layer that lives before auto-merge, sized for the much larger pool of PRs auto-merge's safety policy intentionally skips.)

Today is ${today}. The open-PR queue on aaronjmars/aeon has crossed the threshold where a human reviewer working alone falls behind: yesterday (June 1) eighteen PRs were merged in a single 37-minute Monday catch-up window, but on every prior weekend day they stacked up untouched. As community skill packs become the primary contribution model and external contributors keep landing skill PRs every other day, the queue's steady-state size will keep climbing — skill-scan evaluates one inbound skill PR at a time, but no skill answers the operator's actual morning question: "of the N open PRs right now, which N1 can I merge in one click and which N2 need real review?"

This branch is that answer. It surveys every open PR on a target repo, categorises each by the files it touches, runs scripts/skill-scan.sh against every changed SKILL.md (same scanner skill-scan reuses verbatim), and emits one structured Telegram digest with four risk buckets sorted by PR age. The operator can fire-and-forget the FAST_TRACK bucket, glance at SKILL_PASS, and budget real attention for INFRA_REVIEW + SKILL_WARN_OR_BLOCK + CORE_REVIEW.

Read memory/MEMORY.md for context. Read the last 8 days of memory/logs/ for prior-run context (skip if dispatched). Read soul/SOUL.md + soul/STYLE.md if populated to match voice in the notification.

Why a separate branch from pr-triage / skill-scan / auto-merge

The four compose. pr-triage runs once per PR open; skill-scan runs on demand per skill PR; auto-merge runs against the bot subset; this survey branch is the morning brief over everything else — the open backlog the operator still has to think about. Building a fifth verdict layer into any of the existing three would either bloat their per-PR cost or skip the operator-overview question entirely.

Inputs

No new secrets. GitHub access via gh CLI (GH_TOKEN) per CLAUDE.md.

Writes:

• output/articles/pr-merge-${today}.md — full digest with one row per open PR, sortable by bucket + age (every non-error run, including QUIET)
• memory/topics/pr-merge-state.json — prior-run snapshot (per-PR bucket + first_seen date + last_head_sha, used to suppress re-notification on the same head SHA)
• memory/logs/${today}.md — one log block per run
• Notification via ./notify — only when ≥1 new PR appeared in a non-FAST_TRACK bucket since the last run, or a SKILL_BLOCK / CORE_REVIEW PR is present and operator has not been notified about it on this head SHA yet, or it's the first (baseline) run (see step 8)

Steps

0. Bootstrap

mkdir -p memory/topics output/articles
[ -f memory/topics/pr-merge-state.json ] || cat > memory/topics/pr-merge-state.json <<'EOF'
{"last_run":null,"last_status":null,"last_repo":null,"prs":{}}
EOF

If jq empty fails on the state file (corrupt JSON from an aborted write), back it up to .bak, reset to the empty template, and tag the run STATE_CORRUPT. Continue — a fresh state file means re-notifying every currently-open PR as "new" on this one run, which is the safer post-corruption outcome than silently skipping a SKILL_BLOCK PR the operator hadn't seen yet.

prs is a map keyed by <owner>/<repo>#<number>: {bucket, scan_verdict, head_sha, first_seen, last_notified_head_sha, age_days, author, draft}. Cap to 200 most-recent entries (LRU by first_seen) so a long-lived state file can't grow unbounded.

1. Parse var (survey grammar)

The --survey/survey token has already been consumed by the shared preamble. Parse the remaining tokens:

• Tokens: dry-run, anything matching ^[A-Za-z0-9_.-]+/[A-Za-z0-9_.-]+$ (treated as REPO_OVERRIDE), anything else.
• If any unknown token is present → log PR_MERGE_QUEUE_BAD_VAR: ${var} and exit (no writes, no notify).
• DRY_RUN=yes if the dry-run token is present, else no (execute).
• TARGET_REPO=${REPO_OVERRIDE:-aaronjmars/aeon}.

2. Pull the open-PR list

gh api "repos/${TARGET_REPO}/pulls?state=open&per_page=100" --paginate \
  --jq '[.[] | {number, title, user_login: .user.login, draft, base_ref: .base.ref, head_sha: .head.sha, created_at, updated_at, mergeable_state, labels: [.labels[].name]}]' \
  > /tmp/pr-merge-prs.json

If gh api fails (non-zero exit) → log PR_MERGE_QUEUE_API_FAIL: pulls list, write a one-line notification (pr-review --survey: GitHub API failed listing open PRs for ${TARGET_REPO}), exit API_FAIL. The pulls endpoint is the floor — if it can't be read, every downstream step is meaningless.

Empty list → write the article + state with all buckets at zero, log PR_MERGE_QUEUE_EMPTY, skip notify (an empty queue is not news), exit EMPTY.

Bot/trusted-author short-circuit: build TRUSTED_LOGINS from the ## Trusted Authors section of memory/watched-repos.md (one login per - bullet). Always include the bot logins dependabot[bot], renovate[bot], github-actions[bot], plus aeonframework and aaronjmars. PRs whose user_login is in this set get categorised TRUSTED_AUTHOR and bypass the file-bucket logic in step 4 (auto-merge handles them; surfacing them in the operator's risk view would only add noise).

3. Per-PR files fetch

For each PR (skip drafts and skip TRUSTED_AUTHOR PRs in this step — drafts are signals the contributor isn't done; trusted-author PRs route through auto-merge):

gh api "repos/${TARGET_REPO}/pulls/${PR_NUMBER}/files?per_page=100" --paginate \
  --jq '[.[] | {filename, status}]' \
  > "/tmp/pr-merge-files-${PR_NUMBER}.json"

If the files endpoint fails for a single PR → tag the PR bucket=UNKNOWN, scan_verdict=files_api_fail, skip its scan but keep it in the digest as an UNKNOWN row at the bottom (silently dropping a PR from a triage digest is the failure mode this skill exists to prevent). Continue to the next PR.

4. Bucket by touched files

Apply this rubric in order (first match wins). The rubric is conservative — when in doubt, escalate, never the other way.

Note: CORE_REVIEW takes precedence over INFRA_REVIEW takes precedence over SKILL_*. A PR that touches both aeon.yml AND a skills/*/SKILL.md is CORE_REVIEW, not SKILL_PASS — the executor-config change is the higher-blast-radius signal and the digest should reflect that. The "first match wins" ordering in the rubric encodes this.

5. Per-skill security scan (only for PRs that touched at least one skills/*/SKILL.md)

For each SKILL.md path changed in such a PR:

gh api "repos/${TARGET_REPO}/contents/${PATH}?ref=${HEAD_SHA}" \
  --jq '.content' | base64 -d > "/tmp/pr-merge-scan-${PR_NUMBER}-$(echo ${PATH} | tr '/' '_').md"
bash scripts/skill-scan.sh "/tmp/pr-merge-scan-${PR_NUMBER}-$(echo ${PATH} | tr '/' '_').md" > "/tmp/pr-merge-scan-${PR_NUMBER}.out" 2>&1

The scan output's first HIGH/MEDIUM/PASS verdict line is taken as the per-file verdict. PR-level verdict is the worst across all scanned files: any HIGH → BLOCK; otherwise any MEDIUM → WARN; otherwise PASS.

If scan.sh itself errors (missing, non-executable, non-zero exit without a verdict line) → tag the PR scan_verdict=scan_error, bucket UNKNOWN, do not infer PASS from silence. The scanner is the same path skill-scan relies on; a broken scanner is a fleet-wide problem the operator must know about.

A PR can touch skills/*/SKILL.md AND skills/*/scan.sh AND aeon.yml at the same time. The bucket precedence in step 4 routes it to CORE_REVIEW regardless of the scan verdict — but still run the scan and record the verdict in the state file. A CORE_REVIEW PR with a BLOCK scan verdict deserves a second escalation line in the notification.

6. Age, label, and metadata

For each PR:

• age_days = floor((now - created_at) / 86400)
• updated_age_days = floor((now - updated_at) / 86400) (used as a tie-breaker when sorting — a PR last updated 8 days ago beats one updated yesterday for review urgency)
• has_changes_requested = mergeable_state == "blocked" AND any label in {changes-requested, needs-revision}
• labels = the labels array from the PR list response

Within each bucket, sort by age_days DESC, then updated_age_days DESC, then number ASC. The oldest unreviewed PR in each bucket is the operator's highest-leverage merge click.

7. Write the article

Overwrite output/articles/pr-merge-${today}.md:

# PR Merge Queue — ${TARGET_REPO} — ${today}

*Open PRs surveyed: N · Drafts skipped: D · Trusted-author PRs (auto-merge handles): T*

## FAST_TRACK ({fast_track_count})

Docs / asset / data-only. Safe to merge on title + a glance.

| # | Title | Author | Age | Labels |
|---|-------|--------|-----|--------|
| ... |

## SKILL_PASS ({skill_pass_count})

Skill PR; security scan PASS on every changed `SKILL.md`.

| # | Title | Author | Age | Scan |
|---|-------|--------|-----|------|

## INFRA_REVIEW ({infra_review_count})

Build / CI / dependency-surface change. Requires operator eyes.

| # | Title | Author | Age | Files touched |
|---|-------|--------|-----|---------------|

## SKILL_WARN_OR_BLOCK ({skill_warn_count})

Skill PR with security scan WARN or BLOCK. **Do not merge without resolving the finding.**

| # | Title | Author | Age | Scan | Verdict |
|---|-------|--------|-----|------|---------|

## CORE_REVIEW ({core_review_count})

Touches the runtime executor or `aeon.yml`. **Highest blast radius — review carefully.**

| # | Title | Author | Age | Files touched | Scan (if any) |
|---|-------|--------|-----|---------------|---------------|

## UNKNOWN ({unknown_count})

Files endpoint failed OR matched no rule. Operator: glance and re-run.

| # | Title | Author | Age | Reason |
|---|-------|--------|-----|--------|

## TRUSTED_AUTHOR ({trusted_count})

Routed to `auto-merge`. Listed here for visibility only.

| # | Title | Author | Age |
|---|-------|--------|-----|

---
*Generated by `pr-review --survey`. Bucket precedence: CORE_REVIEW > INFRA_REVIEW > SKILL_WARN_OR_BLOCK > SKILL_PASS > FAST_TRACK > UNKNOWN. Run again with `--survey dry-run` to refresh without sending a notification.*

PR rows: | #N | [title] | @author | Nd | [labels-or-files-or-scan-cell] | — every column ≤80 chars, truncate with … if needed so the markdown table stays narrow enough to render in the dashboard.

8. Decide whether to notify (gated)

Skip notify entirely on BAD_VAR, API_FAIL, EMPTY, DRY_RUN, STATE_CORRUPT.

Otherwise notify only if any of:

1. First (baseline) run — state.prs was empty before this run.
2. New non-FAST_TRACK PR — a PR present this run but not in the prior state, with bucket != FAST_TRACK and bucket != TRUSTED_AUTHOR (a new FAST_TRACK or trusted-author PR is not news; surfacing one would re-create the dependabot-noise pattern other skills work hard to suppress).
3. SKILL_BLOCK or CORE_REVIEW PR with a fresh head SHA — present in this run and either state.prs[k].last_notified_head_sha != head_sha OR state.prs[k] did not exist. (A force-push or rebase reopens the review surface; we re-notify once per head SHA, not per run.)
4. A bucket transitioned worse — e.g. yesterday's SKILL_PASS now scores SKILL_WARN_OR_BLOCK because a new commit added a HIGH finding.

When notifying, set state.prs[k].last_notified_head_sha = head_sha for every PR cited in the notification body, so a noisy queue doesn't re-fire the same line every morning. PRs that drop out (closed/merged) are removed from prs (no retention beyond active queue).

9. Notification format

*PR Merge Queue — ${TARGET_REPO} — ${today}*

{open_count} open · {new_count} new since last run · {trusted_count} trusted (auto-merge)

FAST_TRACK ({n}): #A · #B · #C   ← safe to merge
INFRA_REVIEW ({n}): #D · #E   ← needs eyes
SKILL_PASS ({n}): #F · #G   ← clean scan
SKILL_WARN_OR_BLOCK ({n}): #H ⚠️ HIGH security finding · #I ⚠️ MEDIUM
CORE_REVIEW ({n}): #J ⚠️ touches aeon.yml · #K
{If unknown_count > 0:} UNKNOWN ({n}): #L — files API failed

Oldest in queue: #M (Nd, {bucket})
Full digest: output/articles/pr-merge-${today}.md

Keep under 900 chars. Drop any bucket row whose count is zero. Drop the "oldest in queue" line if open_count == 0. The ⚠️ HIGH security finding / ⚠️ touches aeon.yml annotations are appended only to PRs the operator has not been notified about on this head SHA — repeat-rendering them would dilute the alert signal.

Send via ./notify "$MSG" (single positional argument — the heredoc-built message; aeon's ./notify accepts a positional argument or -f file, this branch uses positional to keep the message inline with the other locals computed in this step, and keeps it under 900 chars).

10. Log (SURVEY branch)

Append to memory/logs/${today}.md under the shared ### pr-review heading with a mode discriminator:

### pr-review
- **Mode**: survey (risk-tiered triage digest)
- **Target repo**: ${TARGET_REPO}
- **Open PRs**: N (drafts skipped: D · trusted-author: T)
- **Buckets**: FAST_TRACK=A · SKILL_PASS=B · INFRA_REVIEW=C · SKILL_WARN_OR_BLOCK=D · CORE_REVIEW=E · UNKNOWN=F
- **New since last run**: G (excluding FAST_TRACK and TRUSTED_AUTHOR)
- **Oldest open PR**: #N (Mt days, bucket={bucket})
- **Scan results**: PASS=P · WARN=W · BLOCK=B · scan_error=E
- **Article**: output/articles/pr-merge-${today}.md
- **Notification**: sent / skipped (gated)
- **Status**: PR_MERGE_QUEUE_OK

Exit taxonomy (SURVEY branch)

PR_MERGE_QUEUE_OK and PR_MERGE_QUEUE_QUIET are the two success states. The split exists so the dashboard can show "ran clean, nothing changed" without overloading the OK row.

Design notes (SURVEY branch — do not edit without reading)

• File-bucket precedence is conservative on purpose. A PR touching both aeon.yml and skills/x/SKILL.md is CORE_REVIEW. Routing it to SKILL_PASS because the scan passed would understate the blast radius — the executor-config change is the dominant signal. First-match-wins in step 4 encodes this; do not "improve" by softening to per-file majority voting.
• Scan errors do NOT default to PASS. scan_error → bucket UNKNOWN. The whole point of the scanner is to catch HIGH findings before merge; a silent PASS on a broken scanner is the failure mode the skill exists to prevent. Same rule skill-scan follows.
• FAST_TRACK is not "ignore." It's the bucket the operator should merge first, not the bucket the operator should skip. A FAST_TRACK PR sitting open for 14 days is still a contributor experience problem; the digest still surfaces it (just without a notification).
• TRUSTED_AUTHOR bypasses every other bucket. A bot PR touching aeon.yml is still a auto-merge-handled PR — listing it under CORE_REVIEW would suggest the operator should look, but auto-merge's policy already gates this. The split keeps the operator's mental model: this digest is the queue you have to think about, not everything that's open.
• Re-notification is gated on head SHA, not on date. A queue that grows by one PR per day shouldn't re-notify yesterday's whole queue every morning. State carries last_notified_head_sha per PR; a force-push or rebase reopens the review surface and is the only thing that re-notifies. Same dedup pattern skill-scan uses.
• No auto-merge action, no PR comments, no labels. This branch is operator-facing only. If a comment-on-PR layer is wanted, that is the default REVIEW branch / pr-triage / skill-scan territory and should grow there.
• memory/watched-repos.md is read for ## Trusted Authors only (in this branch). The branch does not iterate every watched repo by default because the digest format only renders cleanly for one repo at a time and the operator's morning question ("what's safe to merge on aeon today") is per-repo. Multi-repo support is the ${var} override path.

Sandbox note

gh CLI handles GitHub auth internally — use it over raw curl with header expansion the sandbox blocks. Both branches route all outbound GitHub calls through gh / gh api (GH_TOKEN, or GITHUB_TOKEN in CI — provided by the runner, no new secret to provision). No prefetch/postprocess wrapper required. The only other outbound call is ./notify, which is already sandbox-safe.

• REVIEW branch: if gh fails at the repo level, log the error and continue to the next repo. As a last-resort fallback, use WebFetch on the raw PR URL to read the diff.
• SURVEY branch: the pulls list endpoint is the floor (see step 2) — on failure, one-line failure notify + exit API_FAIL. A single PR's files-endpoint failure degrades that PR to UNKNOWN but keeps it in the digest.

No third-party API keys. No on-chain reads. No file writes outside memory/, output/articles/, and /tmp/.