loki-mode 10.5.27 → 10.5.28

package/SKILL.md

@@ -3,7 +3,7 @@ name: loki-mode
 description: An autonomous software factory that knows what it is supposed to deliver, and proves it did. Use when the user says Loki Mode or asks to build, fix or verify software autonomously.
 ---
 
-# Loki Mode v10.5.27
+# Loki Mode v10.5.28
 
 **You are an autonomous agent. You make decisions. You do not ask questions. You do not stop.**
 
@@ -470,4 +470,4 @@ See `CHANGELOG.md` entries [7.5.7], [7.5.8], [7.5.13] for the per-fix list and r
 
 ---
 
-**v10.5.27 | [Autonomi](https://www.autonomi.dev/) flagship product | ~410 lines core**
+**v10.5.28 | [Autonomi](https://www.autonomi.dev/) flagship product | ~410 lines core**

package/VERSION

@@ -1 +1 @@
-10.5.27
+10.5.28

package/dashboard/__init__.py

@@ -7,7 +7,7 @@ Modules:
     control: Session control API (start/stop/pause/resume)
 """
 
-__version__ = "10.5.27"
+__version__ = "10.5.28"
 
 # Expose the control app for easy import
 try:

package/docs/INSTALLATION.md

@@ -2,7 +2,7 @@
 
 The flagship product of [Autonomi](https://www.autonomi.dev/). Loki Mode is a spec-driven autonomous builder with a built-in trust layer that takes any spec to a deployed product and verifies completion with evidence (quality gates plus a completion council), not just a "done" claim. Complete installation instructions for all platforms and use cases.
 
-**Version:** v10.5.27
+**Version:** v10.5.28
 
 ---
 

package/docs/v10/BOARD.md

@@ -611,7 +611,8 @@ Source: ~/git/autonomi-dev/research/2026-09-30-adoption/SWARM-PROMPT-ADOPTION.md
 | A-101 | Receipt canonicalization and verify help (item 1) | see docs/v10/ADOPTION-PLAN.md A-101 | HIGH | first-run-gate assertion per plan card | released@2026-09-30T21:23Z | Wave per plan section 5; one writer per file. Released in v10.5.11 (merge 5ea92836 is an ancestor of v10.5.11). |
 | A-102 | Test-runner detection: node:test plus fixtures for all six runners (item 1) | see plan A-102 | HIGH | first-run-gate assertion per plan card | released@2026-10-01T00:52Z | Wave per plan section 5; one writer per file. Released in v10.5.13 (merge 0990afb8 is an ancestor of v10.5.13). |
 | A-104 | Commit only the fix (items 1 and 2, v10) | see plan A-104 | HIGH | first-run-gate assertion per plan card | released@2026-09-30T21:23Z | Wave per plan section 5; one writer per file. Released in v10.5.11 (merge d86a5817 is an ancestor of v10.5.11). |
-| A-104b | A-104 opus follow-ups: run the unstage reset as `git --literal-pathspecs reset` and check its rc (a path starting with `:(top)` was read as a pathspec and committed); match "manifest changed" per directory in monorepos (a stray lockfile in packages/b rode along with a packages/a manifest change); judge new lockfiles and Wall files against the run base, not HEAD, so files the agent commits during implement are also excluded | loki-ts/src/engine10/stages/seal.ts, loki-ts/tests/engine10/seal.test.ts | HIGH | fixtures for each case red first; seal-then-verify still matches | building@2026-10-01T08:44Z | Source: A-104 opus review (APPROVE with notes). seal.ts writer order: after A-121. |
+| A-104b | A-104 opus follow-ups: run the unstage reset as `git --literal-pathspecs reset` and check its rc (a path starting with `:(top)` was read as a pathspec and committed); match "manifest changed" per directory in monorepos (a stray lockfile in packages/b rode along with a packages/a manifest change); judge new lockfiles and Wall files against the run base, not HEAD, so files the agent commits during implement are also excluded | loki-ts/src/engine10/stages/seal.ts, loki-ts/tests/engine10/seal.test.ts | HIGH | fixtures for each case red first; seal-then-verify still matches | merged@2026-10-01T09:45Z | Source: A-104 opus review (APPROVE with notes). seal.ts writer order: after A-121. |
+| A-104c | A-104b r2 opus follow-up (code reading, not reproduced): a FAILED run now reaches backstopCommit in supervisor.ts (around lines 93-98 and 229), which runs git add -A with hooks off, so a run whose commit stage failed could still commit Wall files, stray lockfiles and the user's pre-run dirty files, and hasDiff could push them as a draft PR (E-67 behaviour). Reproduce first; then make the backstop honor the same exclusions as the commit stage, or skip it when the commit stage itself failed | loki-ts/src/engine10/supervisor.ts, its test | failed-commit run leaves no new commit with Wall files, lockfiles or pre-run dirty files, and opens no PR | HIGH | ready@2026-10-01T09:45Z | none |
 | A-120 | One signing key, auto-generated (item 3) | see plan A-120 | HIGH | first-run-gate assertion per plan card | released@2026-10-01T00:52Z | Wave per plan section 5; one writer per file. Released in v10.5.13 (merge 1e1417b9 is an ancestor of v10.5.13). |
 | A-132 | Legacy: no HANDOFF.md, USAGE.md or lockfile in a fix commit (item 2) | see plan A-132 | MEDIUM | first-run-gate assertion per plan card | released@2026-10-01T00:52Z | Wave per plan section 5; one writer per file. Released in v10.5.13 (merge 297c2d1f is an ancestor of v10.5.13). |
 | A-133 | Legacy: skip Council evidence probes on a trivial diff (item 2) | see plan A-133 | HIGH | first-run-gate assertion per plan card | merged@2026-09-30T21:08Z | Wave per plan section 5; one writer per file. |

package/docs/v10/PROGRESS.md

@@ -1807,3 +1807,25 @@ The swarm is stopped: .loki/V10-STOP exists, the loop has no wake-ups, and the c
 - promote run 36837313500 (08:35Z) failed and held latest. Leg 1 (bun): all checks PASS on the installed 10.5.25 (Loki 10 start line, 7 lines, signed receipt, skip-bare-verify rc=4). Leg 2 (no bun): every legacy check PASS (verify ok, signed, digest, 13/15 lines, skipped target rc=3 NOT VERIFIED) except legacy-fallback-line: the gate matched the pre-S2 message text. Product behaviour on both machines is correct; the gate pattern was stale.
 - Fix 4b290f31: both fallback patterns match bin/loki's current line; the gate test asserts gate and bin/loki carry the same literal (mutation red). test-first-run-gate 78/0, gate --stub rc=0. train/52 in CI with E-161b and E-169; on green cut 10.5.26, whose promote runs the fixed gate.
 - E-170 measured (slim 30,280 KB vs full 407,656 KB); slim-path docs held for the plain-line evidence.
+
+## 2026-10-01T09:15Z v10.5.26 PROMOTED to latest through both gate legs (D48 flip reaches plain npm users)
+- v10.5.26 25246946: npm gitHead 25246946c541 matches; promote run 36841179022 at 09:12Z completed success; dist-tags latest=10.5.26, next=10.5.26; release body 9 lines. First promote since 10.5.20: leg 1 with bun (Loki 10 checks) and leg 2 with no bun (installed --omit=optional, bun off PATH, legacy checks plus the exact fallback line) both passed.
+- Shipped in 10.5.26: gate text fix 4b290f31 (plus a gate/bin/loki literal-sync test), E-161b, E-169. 10.5.25 (bundled bun 1.4.2, resolver, two-leg gate) is now on latest through 10.5.26.
+- Merged since: E-153 (pre-commit foreign-author guard, main checkout), A-04b (loki-seal counter fixes, 34/0), A-134c plus r2 (provider stdin from /dev/null incl. cline; TS run.sh tests 652/0 before and after with a real install). train/53 6b1ade3a in CI.
+- A-104b (Seal: literal-pathspec reset, per-directory lockfiles, judge against baseSha) built, 935/0 engine10; in opus review because item 3 changes what Seal judges.
+
+## 2026-10-01T09:19Z A-104b REJECT (opus): pre-existing moat gap, failed commit stage still seals VERIFIED
+- Repro (opus): agent commits src.js during implement, leaves a.txt uncommitted; a lockfile recorded at intake as ../outside.txt makes the new literal-pathspecs reset fail, so the commit stage returns "failed: git reset failed"; machine.ts still advances to seal and seal.ts never checks commit success, so the receipt reads VERIFIED with a.txt left modified. Predates A-104b; the new rc check never stopped Seal.
+- Also: empty or unknown ctx.baseSha does not fail closed (nothing filtered, junk committed to the run branch; Seal then reads FAILED). Production always sets baseSha (worker.ts:47).
+- Held up: judging against baseSha never drops a real source change from the receipt diff (verifyReceipt and verifyMain VERIFIED rc 0 on the fixture); monorepo per-directory lockfiles correct; suites 94/0, gate --stub 0 failed.
+- A-104b r2 building: machine.ts routes a failed commit to the failure path; seal.ts refuses VERIFIED when commit did not complete; the commit stage fails on empty or unresolvable baseSha or a failed diff. Red-first tests plus the full moat suite.
+
+## 2026-10-01T09:32Z D50 (top engineering priority): loss classification; 3-rep rerun started
+- Baseline artifacts: ~/loki-ci-logs/d50-v10-sonnet and d50-raw-sonnet (results.jsonl, logs/<task>.<arm>.<run>/arm_stdout.log); harness eval/loki10/run.sh at c5eaddb0.
+- Loki+sonnet losses vs raw sonnet (haiku analyst, from arm_stdout.log):
+  - aiq-52-searchbar: dirty-tree refusal ("M frontend/package-lock.json", 0.7s). The baseline ran before E-164 (7b0ba412 merged after c5eaddb0), so this class may be recovered.
+  - pub-click-2877: false ALREADY_SATISFIED (no PR; hidden tests fail). Real defect in the already-done check.
+  - pub-humanize-174: implement ran 95s and produced an empty diff with no already_done marker; FAILED, no PR. Raw sonnet completed it.
+  - pub-humanize-333: PR opened and hidden tests failed; verify did not catch the wrong fix. Raw sonnet completed it.
+  - pub-click-3059: raw sonnet failed it too, so it is not a Loki-specific loss.
+- Rerun (b) started 2026-10-01T09:32Z: the 4 Loki-specific tasks x 3 reps, arms v10 and raw-claude, model claude-sonnet-5, at main 575ddebcf; out ~/loki-ci-logs/d50-rerun-*. No lift number is cited until it finishes. Then fix the top cause first and record a per-model stage profile with a lift row.