Skip to content

Loki Mode

v7.116.0 Bugfix

This release fixes issues for SREs watching stability and regressions.

✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

ai-agents aider anthropic autonomous ci-cd claude
+13 more
cline code-review devops gemini github-action github-issues loki-mode multi-agent openai-codex openapi pull-request-review sdlc spec-driven-development

Summary

AI summary

Fixed false-negative test verification for node --test runners and corrected double .loki directory paths in state markers.

Full changelog

Trust-core fixes surfaced by a live calibration build (RED/GREEN, council-gated)

A real bounded engine build (the shipped v7.115 artifact against a 3-item PRD)
produced a genuinely-working, fully-tested deliverable, and surfaced two real
defects that no fixture caught. Both are fixed here with RED/GREEN tests.

  • TRUST DEFECT: passing tests were recorded as tests: not_run, so
    genuinely-correct, fully-tested work got a NOT VERIFIED headline (false
    negative).
    The test-runner detection did not recognize Node's built-in
    node --test runner (stable since Node 18, runs *.test.{js,mjs,cjs} with
    zero config and NO package.json). A deliverable of slug.js + slug.test.js
    (9/9 passing) fell through to runner:none / status:not_run / verification_gap:source_without_tests, so the proof honestly derived NOT
    VERIFIED -- a false negative that is as corrosive to a trust-as-product engine
    as fake-green. The detection lives in BOTH paths (the in-loop test gate in
    autonomy/run.sh AND the loki verify gate in autonomy/verify.sh); both are
    now fixed. A node --test branch was added as a FALLBACK below every explicit
    path (a package.json declaring vitest/jest/mocha still selects that runner):
    it fires only when node is on PATH and matching test files exist, records
    the REAL result (runner="node-test", pass from exit code, parsed counts), and
    a FAILING suite records pass:false (never swallowed). Source without test
    files still records source_without_tests (no over-fire); no node still falls
    through to the honest inconclusive path (never fabricates a pass). Test:
    tests/test-node-test-detection.sh (9/9 incl. both-paths + failing-honesty +
    jest-no-regression cases).
  • BUG: double-.loki path when writing the COMPLETED / orchestrator state
    markers.
    autonomy/loki sets LOKI_DIR="${LOKI_DIR:-.loki}", and two sites
    in autonomy/run.sh defined loki_dir with the /.loki OUTSIDE the :-
    default ("${LOKI_DIR:-${TARGET_DIR:-.}}/.loki"), so a set LOKI_DIR of .loki
    became .loki/.loki -- the build failed with
    .loki/.loki/COMPLETED: No such file or directory. Both sites
    (_advance_current_phase + main) are corrected to the sibling form
    "${LOKI_DIR:-${TARGET_DIR:-.}/.loki}" (the /.loki inside the default), so a
    set LOKI_DIR is used as-is. Behavior when LOKI_DIR is unset is unchanged. An
    exhaustive repo-wide audit confirms these were the only two buggy sites. Test:
    tests/test-loki-dir-double-path.sh (11/11).

Both tests are wired into run-all-tests.sh + local-ci.sh. No behavior change to
any shipping command surface beyond correctly recording a Node test run and
writing the marker to the right path.

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

Share this release

Track Loki Mode

Get notified when new releases ship.

Sign up free

About Loki Mode

Multi-agent autonomous SDLC framework. Spec to deployed app. PRD, GitHub issue, OpenAPI/JSON/YAML, or one-line brief. 5 AI providers, 11 quality gates.

All releases →

Related context

Beta — feedback welcome: [email protected]