@feigi/fleet-ctl 3.21.7 → 3.21.9

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@feigi/fleet-ctl",
-  "version": "3.21.7",
+  "version": "3.21.9",
   "description": "Agent fleet: run-team controller, merge bot, PR reviewer, and the ticket pipeline they share",
   "license": "Apache-2.0",
   "repository": {

package/scripts/finisher-own-tree-prose.test.mjs

@@ -205,12 +205,13 @@ test("duty 1 says worktree-audit.sh takes no argument and audits every worktree"
 // equal alternatives, then halted on "dirty or diverged" — and neither can
 // see diverged. Porcelain is empty over a clean tree one commit past the pin,
 // and the audit's `ahead` is counted against `origin/main`
-// (worktree-audit.sh's `base=${BASE_REF:-origin/main}`), which a PR worktree
-// is always ahead of. The only divergence check is the cause block's
-// head-equality against the dispatch pin. And "find this worktree's row" gave
-// no rule for a row that cannot answer, or for no row at all. Each pin below
-// holds ONE sentence (`sentences()`, #1940), so a clause deleted from inside
-// the duty reds its own pin rather than being satisfied by a neighbour.
+// (worktree-audit.sh's `base=${BASE_REF:-origin/main}`), which a healthy
+// worktree on an open, unmerged PR is ahead of too. The only divergence check
+// is the cause block's head-equality against the dispatch pin. And "find this
+// worktree's row" gave no rule for a row that cannot answer, or for no row at
+// all. Each pin below holds ONE sentence (`sentences()`, #1940), so a clause
+// deleted from inside the duty reds its own pin rather than being satisfied by
+// a neighbour.
 test("duty 1 names the instrument for each check, and halts on a row that cannot answer", () => {
   const said = (re) => sentences(duty1Text()).some((s) => re.test(s));
   assert.ok(
@@ -225,6 +226,19 @@ test("duty 1 names the instrument for each check, and halts on a row that cannot
     said(/row's `ahead`.*counted against `origin\/main`.*cannot answer/),
     "duty 1 no longer says the audit's `ahead` is counted against origin/main — a finisher reading it as divergence sees every PR worktree as diverged, or learns to ignore it",
   );
+  // The reason `ahead` cannot answer holds for the PR the finisher runs on: an
+  // open, unmerged one. It is not true of every PR worktree — after a merge
+  // commit into a fetched origin/main the branch's `ahead` reads 0 — so the
+  // claim is scoped, never "always", and the scope sits in the same sentence.
+  assert.ok(
+    said(/row's `ahead`.*counted against `origin\/main`.*open, unmerged PR.*cannot answer/),
+    "duty 1 no longer scopes the `ahead` reason to an open, unmerged PR's worktree — after a merge into a fetched origin/main a PR worktree's `ahead` reads 0, so an unscoped claim is false",
+  );
+  assert.doesNotMatch(
+    flat(RUN_TEAM),
+    /\balways\b[^.]{0,20}\bahead\b/i,
+    "the run-team skill states an unconditional lead over origin/main again — a PR worktree's `ahead` reads 0 once its merge commit is in a fetched origin/main",
+  );
   // The three non-clean row shapes, each halting and each named. Unreadable
   // and missing share `readable:false` and differ only in their counts, so
   // each is pinned with its count shape and stderr line — dropping one leaves

package/skills/run-team/SKILL.md

@@ -2578,10 +2578,10 @@ a minute apart showed *different* mutants, so a member's report and any single
    the worktree unreadable. **Diverged:** the worktree's `HEAD` against your
    dispatch pin — the cause block's head-equality check below. Neither
    instrument above answers it: `git status --porcelain` reports no commits,
-   and the row's `ahead` is counted against `origin/main`, which a PR worktree
-   is always ahead of, so it cannot answer this. Three row shapes halt too,
-   each named in the report with the script's stderr line quoted where it
-   prints one: `readable:false` with `null` counts
+   and the row's `ahead` is counted against `origin/main`, which a healthy
+   worktree on an open, unmerged PR is ahead of too, so it cannot answer this.
+   Three row shapes halt too, each named in the report with the script's
+   stderr line quoted where it prints one: `readable:false` with `null` counts
    (`UNREADABLE: <wt> (<cause>)`) is *worktree unreadable*, `readable:false`
    with zero counts (`MISSING on disk: <wt>`) is *worktree missing*, and no
    row for this worktree at all is *worktree absent*. `git status` is no