omnius 1.0.691 → 1.0.693

package/docs/DISCOVERY.json

@@ -33864,6 +33864,84 @@
         }
       ]
     },
+    {
+      "id": "guide.work-orders-runtime-health-remediation-wo-22-autocomplete-footer-geometry-uppercase",
+      "kind": "guide",
+      "title": "WO-22: Autocomplete Footer Geometry Integrity",
+      "summary": "Scope: CLI TUI footer, DirectInput redraw scheduling, active content writes",
+      "keywords": [
+        "work",
+        "orders",
+        "runtime",
+        "health",
+        "remediation",
+        "WO",
+        "22",
+        "autocomplete",
+        "footer",
+        "geometry",
+        "md"
+      ],
+      "maturity": "internal",
+      "audiences": [
+        "maintainer",
+        "large-context-agent"
+      ],
+      "layer": "documentation",
+      "interfaces": [
+        {
+          "type": "file",
+          "target": "docs/work-orders/runtime-health-remediation/WO-22-autocomplete-footer-geometry.md"
+        }
+      ],
+      "references": [
+        {
+          "type": "documentation",
+          "target": "docs/work-orders/runtime-health-remediation/WO-22-autocomplete-footer-geometry.md",
+          "relation": "canonical-artifact"
+        }
+      ]
+    },
+    {
+      "id": "guide.work-orders-runtime-health-remediation-wo-23-long-horizon-control-plane-reconciliation-uppercase",
+      "kind": "guide",
+      "title": "WO-23: Long-horizon control-plane reconciliation",
+      "summary": "The restored telegramtest run tui-3869250-mtlvp324-1788468819792-1 continued a large pentest-tooling frontend task across a process restart. It made real source mutations and advanced from the first implementation group to todo-p2a. The run was not dead and its model context had ample nominal capacity.",
+      "keywords": [
+        "work",
+        "orders",
+        "runtime",
+        "health",
+        "remediation",
+        "WO",
+        "23",
+        "long",
+        "horizon",
+        "control",
+        "plane",
+        "reconciliation",
+        "md"
+      ],
+      "maturity": "internal",
+      "audiences": [
+        "maintainer",
+        "large-context-agent"
+      ],
+      "layer": "documentation",
+      "interfaces": [
+        {
+          "type": "file",
+          "target": "docs/work-orders/runtime-health-remediation/WO-23-long-horizon-control-plane-reconciliation.md"
+        }
+      ],
+      "references": [
+        {
+          "type": "documentation",
+          "target": "docs/work-orders/runtime-health-remediation/WO-23-long-horizon-control-plane-reconciliation.md",
+          "relation": "canonical-artifact"
+        }
+      ]
+    },
     {
       "id": "guide.work-orders-telegram-dmn-wo-22-dmn-outreach-and-learning-uppercase",
       "kind": "guide",

package/docs/DISCOVERY.md

@@ -560,6 +560,8 @@ Daemon equivalents are `GET /v1/discovery/bootstrap`, `GET /v1/discovery?q=<inte
 | `guide.work-orders-runtime-health-remediation-wo-19-clean-build-reproducibility-uppercase` | WO-19: Clean Build and Optional-Dependency Reproducibility | Status: deterministic acceptance complete Risk: medium Depends on: WO-00 |
 | `guide.work-orders-runtime-health-remediation-wo-20-exact-session-continuation-uppercase` | WO-20: Exact session continuation across exit and restart | On 2026-09-03, the telegramtest TUI displayed the latest pentest task during startup, but the first restored model context combined a completed breach.py handoff with an unrelated, older page.tsx todo tree. The user then entered continue, and Omnius continued the stale tree instead of the task that was active before /quit. |
 | `guide.work-orders-runtime-health-remediation-wo-21-task-convergence-and-todo-scope-uppercase` | WO-21: Task convergence and todo scope | On 2026-09-03, a user reported that Omnius remained at task 2/8 for about 40 turns. The report did not include a screenshot or run artifact, so the exact label remains unverified. Omnius has two similar counters: |
+| `guide.work-orders-runtime-health-remediation-wo-22-autocomplete-footer-geometry-uppercase` | WO-22: Autocomplete Footer Geometry Integrity | Scope: CLI TUI footer, DirectInput redraw scheduling, active content writes |
+| `guide.work-orders-runtime-health-remediation-wo-23-long-horizon-control-plane-reconciliation-uppercase` | WO-23: Long-horizon control-plane reconciliation | The restored telegramtest run tui-3869250-mtlvp324-1788468819792-1 continued a large pentest-tooling frontend task across a process restart. It made real source mutations and advanced from the first implementation group to todo-p2a. The run was not dead and its model context had ample nominal capacity. |
 | `guide.work-orders-telegram-dmn-wo-22-dmn-outreach-and-learning-uppercase` | WO-22: DMN outreach, DM sharing, and outcome learning | On 2026-09-03 at 17:07 PDT the bot posted in the OMNIUS group without being addressed. The operator asked whether this was self-induced reflection. |
 | `guide.work-orders-telegram-dropbear-context-rca-workorder` | Telegram Dropbear Context Engineering RCA Work Order | Observed run: /home/roko/Documents/Projects/Adjacent/telegramtest/.omnius, run id 1782873796963-i5r7mv. |
 | `guide.work-orders-wo-am-gaps-uppercase` | Associative Memory Gap Work Orders | Generated: 2026-04-13 Source: Deep audit of multimodal associative memory systems Status: READY FOR IMPLEMENTATION |

package/docs/work-orders/runtime-health-remediation/TRACKER.md

@@ -254,6 +254,23 @@ deterministically repaired P0 or P1 defect.
   publish-tree changes.
 - [x] Record WO-19 acceptance evidence in `WO-19-CLEAN-BUILD-EVIDENCE.md`.
 
+## WO-23: long-horizon control-plane reconciliation
+
+- [x] Reconcile typed failure outcomes across retry, workboard, and durable
+  failure memory.
+- [x] Project all unresolved todo leaves and active failure obligations into
+  integration closure.
+- [x] Keep lexical discovery candidates advisory until evidence adopts them.
+- [x] Compact oversized raw discovery independently of total provider context
+  capacity and restore validated canonical receipt handles.
+- [x] Activate nested Git tracking only from explicit in-workspace target paths
+  and reject ambiguous or escaped repositories.
+- [x] Consolidate Telegram `edited_message` revisions before UI output,
+  cognition, attention inference, or queue admission.
+- [x] Pass deterministic regression tests, affected typechecks, and the
+  workspace build.
+- [x] Commit and push the scoped repair to `main`.
+
 ## Program completion gate
 
 - [x] All repository-controlled deterministic WO acceptance sections pass.
@@ -320,6 +337,15 @@ deterministically repaired P0 or P1 defect.
   - [x] Latch an advisory convergence review when the leaf frontier is static.
   - [x] Make loop, reminder, budget, and workboard progress semantics truthful.
   - [x] Pass deterministic race, stasis, display, typecheck, and build tests.
+- [x] WO-22 autocomplete footer geometry uses the last painted layout as its
+  terminal authority and clears all released rows during idle and active-stream
+  contraction.
+  - [x] Trace the deferred DirectInput repaint and content-write ordering race.
+  - [x] Confirm existing tests do not model final terminal cells in this path.
+  - [x] Unify idle and streaming geometry transitions around deterministic
+    replay.
+  - [x] Cover precommitted multi-match to fewer-match and zero-match contraction.
+  - [x] Pass focused footer tests, CLI typecheck, and CLI build.
 - [x] A recovered pause no longer strands its session. Submitting a new task
   retires the superseded generation instead of refusing every later prompt,
   while unreconciled external effects still fence admission.

package/docs/work-orders/runtime-health-remediation/WO-22-autocomplete-footer-geometry.md

@@ -0,0 +1,93 @@
+# WO-22: Autocomplete Footer Geometry Integrity
+
+**Status:** complete
+
+**Observed:** 2026-09-03
+
+**Scope:** CLI TUI footer, DirectInput redraw scheduling, active content writes
+
+## Failure
+
+When slash-command autocomplete expands and then contracts, rows from the old
+footer can remain visible above the current input box. The captured failure
+occurred while `file_read` content was active. A complete old input box and old
+suggestion rows remained in the content viewport after the live footer moved
+down.
+
+## Root cause
+
+DirectInput emits a change and schedules footer painting with `setImmediate`.
+If content output starts before that callback, `beginContentWrite()` recomputes
+and stores the new footer height. The later renderer reconstructs the supposed
+old footer top from that already-mutated height. It therefore observes no
+height change and skips the transition clear.
+
+The terminal's last painted geometry and the logical layout are separate
+states. `_lastFooterPaintTop` is the authority for the former. A mutable
+`_currentFooterHeight` cannot be used to infer what remains on screen.
+
+The streaming renderer also uses incremental line-feed and reverse-index
+compensation when the footer moves. That path depends on terminal scroll-region
+semantics and does not deterministically rebuild released rows from Omnius's
+scrollback model.
+
+## Invariants
+
+- Every footer renderer compares the new top with the last successfully painted
+  footer top.
+- A changed logical height or changed painted top is a geometry transition.
+- Geometry transitions update DECSTBM, clear the complete old/new footer union,
+  replay content from the virtual scrollback, repaint tasks, and then paint the
+  new footer.
+- Active content writes do not use line feed, reverse index, or cursor-offset
+  guesses to compensate for footer growth or shrinkage.
+- After a geometry transition, token chunks enter the bounded 33 ms replay
+  scheduler. They do not cause synchronous full-viewport repainting per token.
+- Replay-consumed stdout writes preserve callback completion, and stream close
+  paints any final coalesced line before returning ownership to the idle footer.
+- Cursor-preserving animation and metrics frames cannot accept a precomputed
+  footer position before the transition is cleared.
+- Footer chrome writes bypass the buffered content layer.
+- The native streaming cursor is restored by content replay, while the input
+  cursor remains owned by the idle path.
+- Repeated expansion and contraction is idempotent and cannot accumulate rows.
+
+## Work items
+
+- [x] Trace DirectInput change ordering and scheduled redraw ownership.
+- [x] Trace `beginContentWrite()` layout mutation and terminal paint ordering.
+- [x] Confirm the idle source-level regression test does not cover active writes.
+- [x] Confirm the PTY helper does not interpret final terminal cells.
+- [x] Make last-painted footer geometry authoritative in the idle renderer.
+- [x] Route active-stream geometry transitions through deterministic replay.
+- [x] Preserve the direct terminal bypass for chrome paints.
+- [x] Add a test for a precommitted logical shrink before idle paint.
+- [x] Add a test for a precommitted logical shrink during active content output.
+- [x] Cover multi-match to fewer-match and multi-match to zero-match transitions.
+- [x] Prove no reverse-index compensation remains in the streaming transition.
+- [x] Run the focused footer and DirectInput suite.
+- [x] Run CLI typecheck and build.
+- [x] Inspect the scoped diff and record acceptance evidence.
+
+## Acceptance evidence
+
+- `StatusBar.paintedFooterTop()` uses `_lastFooterPaintTop` as physical screen
+  authority, with a logical-height fallback only before the first paint.
+- Both footer render paths treat either a height change or a painted-top change
+  as a geometry transition.
+- Active-stream transitions update DECSTBM, clear the old/new footer union, fill
+  and replay the viewport, refresh tasks, and repaint the footer. Later chunks
+  stay in a coalesced virtual replay until the stream closes.
+- `footer-suggestion-shrink-render.test.ts` executes five-to-one and one-to-zero
+  contraction after precommitting the logical height. Its ANSI framebuffer
+  confirms that prior suggestion text is absent from final terminal cells.
+- The same test executes the five-to-one race with `writeDepth > 0` and confirms
+  the replay path. A following token does not trigger an immediate repaint, its
+  write callback completes, and stream close performs the final replay.
+- A cursor-preserving-frame regression proves that a precomputed logical layout
+  cannot overwrite the last-painted geometry before cleanup.
+- The focused footer/DirectInput suite passed: 6 files, 44 tests.
+- The complete CLI suite passed: 262 files, 2,446 tests.
+- `pnpm --filter omnius typecheck` passed.
+- `pnpm --filter omnius build` passed.
+- No live inference or CUDA workload was used.

package/docs/work-orders/runtime-health-remediation/WO-23-long-horizon-control-plane-reconciliation.md

@@ -0,0 +1,154 @@
+# WO-23: Long-horizon control-plane reconciliation
+
+## Incident
+
+The restored `telegram_test` run
+`tui-3869250-mtlvp324-1788468819792-1` continued a large pentest-tooling
+frontend task across a process restart. It made real source mutations and
+advanced from the first implementation group to `todo-p2a`. The run was not
+dead and its model context had ample nominal capacity.
+
+The live artifacts nevertheless exposed contradictions that make long-horizon
+work slower and less trustworthy:
+
+- the workboard had 34 unresolved cards while integration closure reported
+  `ready`, no gaps, and no blockers;
+- successful corrective actions did not retire several failure cards;
+- todo leaf cards had no evidence even when their declared artifacts were
+  mutated and inspected;
+- restored historical discovery transactions occupied most of the model-visible
+  tool context;
+- a generic task word such as `left` promoted unrelated mesh paths into the
+  affected-path frontier;
+- Git progress stayed permanently disabled because the launch workspace was
+  not a repository even though mutations targeted a nested repository; and
+- repeated `edited_message` updates from a streaming peer bot were displayed
+  and admitted as repeated cognition before logical-input consolidation.
+
+This order repairs Omnius. It does not mutate the pentest project used as the
+field workload.
+
+## Evidence
+
+- Live debug index:
+  `/home/roko/Documents/Projects/Adjacent/telegram_test/.omnius/debug-library/runs/tui-3869250-mtlvp324-1788468819792-1/README.md`
+- Live workboard:
+  `/home/roko/Documents/Projects/Adjacent/telegram_test/.omnius/workboards/tui-3869250-mtlvp324-1788468819792-1/active.json`
+- Live Git projection:
+  `/home/roko/Documents/Projects/Adjacent/telegram_test/.omnius/git-progress/latest.json`
+- Live interruption lifecycle:
+  `/home/roko/Documents/Projects/Adjacent/telegram_test/.omnius/interruptions/tui-3869250-mtlvp324.json`
+- Canonical incident packet:
+  `/home/roko/.codex/attachments/a27b0a6c-c533-49d8-8a4c-818769e24453/pasted-text-1.txt`
+
+## Design invariants
+
+### Failure lifecycle
+
+- [x] Record a failure only after any inline permission retry reaches its final
+      outcome.
+- [x] Never persist a `workboardCardId` unless that card exists.
+- [x] Resolve failures from typed target and receipt evidence, not prose.
+- [x] Include receipt-observed paths when matching a corrective verifier.
+- [x] Retire or supersede the durable obligation when an in-memory failure
+      expires. Expiry is not verification.
+- [x] Surface failed reconciliation as bounded telemetry instead of swallowing
+      it silently.
+
+### Todo and completion truth
+
+- [x] Keep every active todo leaf in the authoritative workboard projection.
+      Apply limits only when rendering.
+- [x] Use canonical path fields for declared-artifact evidence predicates.
+- [x] Bind a mutation or verifier only to an explicit todo/card assignment or
+      an unambiguous normalized artifact overlap.
+- [x] Reject ambiguous todo evidence binding instead of selecting by prose.
+- [x] Give integration closure the current unresolved todo leaves and active
+      project failure obligations.
+- [x] Never render closure as `ready` while either authoritative set remains
+      unresolved.
+
+### Context and ontology
+
+- [x] Classify fused canonical task frames and operational work graphs as
+      working-set context even when they contain a nested current-goal section.
+- [x] Trigger consolidation from raw-discovery pressure independently of the
+      provider's total context capacity.
+- [x] Preserve or reconstruct validated receipt handles across restoration so
+      successful historical discovery does not require full raw transactions.
+- [x] Keep lexical candidate paths advisory.
+- [x] Put only explicit required artifacts, mutations, evidence targets, and
+      adopted claim paths in `affectedPaths`.
+- [x] A single generic task-token overlap must not create an affected path.
+
+### Nested Git progress
+
+- [x] Keep the launch workspace as the safety and `.omnius` state root.
+- [x] Discover Git roots only from canonical tool target or mutation paths.
+- [x] Never recursively select the first nested repository.
+- [x] Recover from an outer `not_repo` result when later typed path evidence
+      identifies one enclosing repository.
+- [x] Reject paths outside the workspace and conflicting repository evidence.
+- [x] Mark post-mutation discovery as late and do not fabricate a pre-mutation
+      clean baseline.
+
+### Telegram edit ingress
+
+- [x] Use `(chat_id, message_id)` as the logical input identity across Telegram
+      `message` and `edited_message` envelopes.
+- [x] Consolidate an edit before conversation observation, TUI `heard` output,
+      queue admission, or attention inference.
+- [x] Replace the queued payload with the newest edit and extend only the
+      bounded quiet window.
+- [x] An edit to an already active or settled peer-bot input must not start a
+      second cognition/reply run.
+- [x] Preserve the latest edited content as transport and conversation evidence.
+- [x] Do not add sender names, text patterns, or forced reply decisions.
+
+## Granular implementation checklist
+
+- [x] Add unresolved todo/failure inputs to `BuildIntegrationClosureInput`.
+- [x] Render bounded todo and failure proof gaps in the closure frontier.
+- [x] Add closure tests for open todos, active failures, and candidate-only
+      paths.
+- [x] Remove advisory candidates from authoritative `affectedPaths`.
+- [x] Add a discovery regression for `left sidebar` versus unrelated
+      `left_*` mesh assets.
+- [x] Correct fused-frame request-budget classification precedence.
+- [x] Correct the equivalent context-dump classification.
+- [x] Add fused-frame accounting tests.
+- [x] Include execution-receipt observed paths in target extraction.
+- [x] Make first observed failures card/store consistent.
+- [x] Reconcile expiry with durable cards without claiming verification.
+- [x] Add failure-resolution error telemetry.
+- [x] Remove the 20-leaf authority cap from todo mirroring.
+- [x] Normalize declared-artifact predicates as paths.
+- [x] Add an exact, ambiguity-safe todo evidence selector.
+- [x] Add todo evidence-binding tests.
+- [x] Add evidence-driven nested Git-root activation.
+- [x] Add nested/multiple/outside-root Git tests.
+- [x] Move edit consolidation before Telegram observation and ingress output.
+- [x] Add repeated peer-bot edit tests for history, queue, and router cardinality.
+- [x] Run affected package tests.
+- [x] Run affected package typechecks.
+- [x] Run the workspace build.
+- [x] Inspect and stage only task-owned files.
+- [x] Commit and push `main` without rewriting history.
+
+## Acceptance
+
+The repair is accepted when deterministic tests prove the invariants above and
+the live evidence can no longer produce these contradictions. Token-generating
+live inference is outside this order unless the mandatory GPU preflight is
+completed first.
+
+## Verification evidence
+
+- `pnpm --filter @omnius/orchestrator typecheck`: passed.
+- `pnpm --filter omnius typecheck`: passed.
+- Focused orchestrator suites: 67 tests passed across 9 files.
+- Telegram transport contract: 11 tests passed.
+- Full `agenticRunner.test.ts`: 175 tests passed.
+- `pnpm -r build`: passed for all 11 buildable workspace projects.
+- No live token-generating inference was used. The mandatory accelerator gate
+  therefore did not apply.

package/npm-shrinkwrap.json

@@ -1,12 +1,12 @@
 {
   "name": "omnius",
-  "version": "1.0.691",
+  "version": "1.0.693",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "omnius",
-      "version": "1.0.691",
+      "version": "1.0.693",
       "bundleDependencies": [
         "image-to-ascii"
       ],
@@ -2011,6 +2011,22 @@
         "node": ">= 8"
       }
     },
+    "node_modules/@peculiar/asn1-asym-key": {
+      "version": "2.9.4",
+      "resolved": "https://registry.npmjs.org/@peculiar/asn1-asym-key/-/asn1-asym-key-2.9.4.tgz",
+      "integrity": "sha512-s7SJfjcXlR3MYMngDTeMLCy3VjGaIxeEy905CQ0j09pDbeYhNBvBGA7r7cAvCZYkVFo9kVg9RAki0K2iUz7SGQ==",
+      "license": "MIT",
+      "dependencies": {
+        "@peculiar/asn1-pkcs8": "^2.9.4",
+        "@peculiar/asn1-schema": "^2.9.4",
+        "@peculiar/asn1-x509": "^2.9.4",
+        "asn1js": "^3.0.10",
+        "tslib": "^2.8.1"
+      },
+      "engines": {
+        "node": ">=14"
+      }
+    },
     "node_modules/@peculiar/asn1-cms": {
       "version": "2.9.4",
       "resolved": "https://registry.npmjs.org/@peculiar/asn1-cms/-/asn1-cms-2.9.4.tgz",
@@ -2167,6 +2183,22 @@
         "node": ">=14"
       }
     },
+    "node_modules/@peculiar/asn1-x509-post-quantum": {
+      "version": "2.9.4",
+      "resolved": "https://registry.npmjs.org/@peculiar/asn1-x509-post-quantum/-/asn1-x509-post-quantum-2.9.4.tgz",
+      "integrity": "sha512-GD0k7BY3dsEIeai7C7bVRPddj/PzZu+sTawRPRxdHKbdy1f9b58WQyk0eFJdr28GShyLVTU8poyi4+bTAZfEtQ==",
+      "license": "MIT",
+      "dependencies": {
+        "@peculiar/asn1-asym-key": "^2.9.4",
+        "@peculiar/asn1-schema": "^2.9.4",
+        "@peculiar/asn1-x509": "^2.9.4",
+        "asn1js": "^3.0.10",
+        "tslib": "^2.8.1"
+      },
+      "engines": {
+        "node": ">=14"
+      }
+    },
     "node_modules/@peculiar/json-schema": {
       "version": "1.1.12",
       "resolved": "https://registry.npmjs.org/@peculiar/json-schema/-/json-schema-1.1.12.tgz",
@@ -2205,18 +2237,19 @@
       }
     },
     "node_modules/@peculiar/x509": {
-      "version": "2.0.0",
-      "resolved": "https://registry.npmjs.org/@peculiar/x509/-/x509-2.0.0.tgz",
-      "integrity": "sha512-r10lkuy6BNfRmyYdRAfgu6dq0HOmyIV2OLhXWE3gDEPBdX1b8miztJVyX/UxWhLwemNyDP3CLZHpDxDwSY0xaA==",
+      "version": "2.1.0",
+      "resolved": "https://registry.npmjs.org/@peculiar/x509/-/x509-2.1.0.tgz",
+      "integrity": "sha512-IYbg1R03CSQGWwl24kGyqrdVtixNSbRaDvBg1r5wyYjTP+VwPQkka1BzTgU5+vxiuwqg04OxdvdJ1xYYFIdUSA==",
       "license": "MIT",
       "dependencies": {
-        "@peculiar/asn1-cms": "^2.6.0",
-        "@peculiar/asn1-csr": "^2.6.0",
-        "@peculiar/asn1-ecc": "^2.6.0",
-        "@peculiar/asn1-pkcs9": "^2.6.0",
-        "@peculiar/asn1-rsa": "^2.6.0",
-        "@peculiar/asn1-schema": "^2.6.0",
-        "@peculiar/asn1-x509": "^2.6.0",
+        "@peculiar/asn1-cms": "^2.9.4",
+        "@peculiar/asn1-csr": "^2.9.4",
+        "@peculiar/asn1-ecc": "^2.9.4",
+        "@peculiar/asn1-pkcs9": "^2.9.4",
+        "@peculiar/asn1-rsa": "^2.9.4",
+        "@peculiar/asn1-schema": "^2.9.4",
+        "@peculiar/asn1-x509": "^2.9.4",
+        "@peculiar/asn1-x509-post-quantum": "^2.9.4",
         "pvtsutils": "^1.3.6",
         "tslib": "^2.8.1",
         "tsyringe": "^4.10.0"
@@ -5128,9 +5161,9 @@
       }
     },
     "node_modules/jose": {
-      "version": "6.2.10",
-      "resolved": "https://registry.npmjs.org/jose/-/jose-6.2.10.tgz",
-      "integrity": "sha512-iiW7J9qRFlGxvCOIBDBDxFePQSn7ZMAnrYGhrrOo6siO/MIqwfyilLR27pkfDgUk+raLuzADS8A3S/KLBisc0g==",
+      "version": "6.2.11",
+      "resolved": "https://registry.npmjs.org/jose/-/jose-6.2.11.tgz",
+      "integrity": "sha512-A5NPn7g8EAzGU3IzRs+Yiq8K5n3ypYS75M5+KKiVHdUexfpWK1kP4ZMq7QnTGDoMj6TJ1dtcEJjW60yZDXS4hg==",
       "license": "MIT",
       "funding": {
         "url": "https://github.com/sponsors/panva"
@@ -7591,9 +7624,9 @@
       }
     },
     "node_modules/undici": {
-      "version": "7.29.0",
-      "resolved": "https://registry.npmjs.org/undici/-/undici-7.29.0.tgz",
-      "integrity": "sha512-IDxfleLmmbSskfWSUATiN1nfn2rDuvnMOqb5CWR92iIfojA0Ud+ulOAAEQ57LPr9rWmsreUyf5lwyao+7GNNVw==",
+      "version": "7.29.1",
+      "resolved": "https://registry.npmjs.org/undici/-/undici-7.29.1.tgz",
+      "integrity": "sha512-RYONW2MeafgYlkVOKYKkA/Ag7BmXqgIWCa8t1m0JcxrQg9pI9lEqRhAOruOBCbAohOa/gkCF+iPi9hrgvTzu6Q==",
       "license": "MIT",
       "engines": {
         "node": ">=20.18.1"

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "omnius",
-  "version": "1.0.691",
+  "version": "1.0.693",
   "description": "AI coding agent powered by open-source models (Ollama/vLLM) — interactive TUI with agentic tool-calling loop",
   "type": "module",
   "main": "./dist/library.js",