Skip to content

armgit / AiSOC

Public Imported from GitHub
0

Open-source AI Security Operations Center: alert fusion, LLM-agent triage, MITRE ATT&CK investigation, and a replayable decision ledger for every agent step. Self-hostable, runs with no API keys, MIT licensed. Ships an MCP server for Claude, Cursor and Continue.

Merge remote-tracking branch 'origin/main' into depth/9.6-native-responder

Beenu Arora57c66ef6334e
diff --cc .github/workflows/ci.yml
index 2225fcae2,31cae7ca8..ce24e26d9
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@@ -701,16 -701,18 +701,31 @@@ jobs
run: python3 scripts/check_cloud_native_collection.py --self-test
- name: Cloud-native collection gate
run: python3 scripts/check_cloud_native_collection.py
++<<<<<<< HEAD
+ # Depth 9.6. Voice input transcribes on the device and sends text, and
+ # "no audio leaves the device" is a claim a comment cannot carry. This
+ # refuses any capture API or binary request body under
+ # apps/mobile/src. Also wired into mobile.yml, which only runs when
+ # that app changes — here it is unconditional, because a claim this
+ # strong should not be graded by a path filter.
+ - name: No audio egress gate — self-test
+ run: python3 scripts/check_no_audio_egress.py --self-test
+ - name: No audio egress gate
+ run: python3 scripts/check_no_audio_egress.py
++=======
+ # Depth 9.2. The probe set behind the recall measurement has to stay
+ # usable: every probe must name an item that is in the corpus, the
+ # probes must stay low-overlap with their answers (a probe written in
+ # the document's own words hands the lexical baseline a score it could
+ # not earn on a real question), and the baseline must still be a real
+ # retriever rather than a stub. The scorer itself runs too: with no
+ # embedding backend on a runner it prints "not measured" and exits 0,
+ # which is the honest row and also proves the refusal path works.
+ - name: Memory recall scorer — self-test
+ run: python3 scripts/score_memory_recall.py --self-test
+ - name: Memory recall scorer
+ run: python3 scripts/score_memory_recall.py
++>>>>>>> origin/main
# One layer in from the name: the profile gate asks whether a connector
# reaches a normalization path, this asks which OCSF class it reaches
# when it gets there. The class uid divided by a thousand is the
diff --cc DEPTH_PROGRESS.md
index d6ef08d53,52b94ffa9..92df71b56
--- a/DEPTH_PROGRESS.md
+++ b/DEPTH_PROGRESS.md
@@@ -467,7 -500,7 +534,11 @@@ invented fields on the first run of th
| Date | Item | What happened |
|---|---|---|
++<<<<<<< HEAD
+| 2026-10-11 | 9.6 | Reproduced: `grep -c 'eas build|gradlew|xcodebuild|expo prebuild' .github/workflows/mobile.yml` returned **0** and the workflow actively *required* the README to say no device build had happened; `grep -rni 'speech|voice' apps/mobile/src apps/mobile/app` returned **0**; and `services/realtime/src/push.ts` sent every record through `webpush.sendNotification`, so a native token was accepted and could never be delivered to. Built the Android APK job, the simulator job, the on-device voice module and the APNs/FCM sender. Negative control: removing the on-device refusal so the cloud recogniser could be used fails two of the eight voice tests. The README guard was inverted rather than deleted — "we added a build so the honesty check is no longer needed" is how an ungated claim comes back. |
++=======
+ | 2026-10-11 | 9.2 | Reproduced: `knowledge_base.py` searched with `to_tsvector`/`plainto_tsquery` and nothing else; `grep -rn qdrant services/api/app/` found only the migration runner and the tenant-deletion sweep, no recall path; and no memory row anywhere carried an author, a source, a scope, an expiry or a state. Built the store of record, the Qdrant index over it, the conflict state machine, the case-closure writer and the recall scorer. Two negative controls recorded in the PR: restoring last-write-wins on a contradiction fails three tests, and restoring vector truncation fails one. The measurement is negative and is published as such — see D11, which also records that it caught the truncation defect. |
++>>>>>>> origin/main
| 2026-10-07 | 0.1 | This file created at base commit `1b8bc2d4`. |
| 2026-10-07 | 0.2 | Every figure re-derived. Two matched exactly (detections, unreachable families); executor arms measured 74 against a captured 73; the cloud/identity/SaaS/code figure measured 395 against a captured 461 on a grouping the plan does not pin, recorded above. Two measurement caveats found: executable and quarantined overlap by 1,724 rules, and the "69% Windows" figure does not reproduce from the index. |
| 2026-10-09 | 4.3 | Reproduced: `check_retention_window.py` on the unmodified tree reported six problems — the lake deleting at 90 days against a 3,650-day cap, no `MODIFY TTL` migration, tiering deleting at 90, and the purge worker calling neither `alerts_under_legal_hold` nor `may_purge` and reporting nothing withheld. Fixed all but the cold-query bullet. Four negative controls recorded in the PR, including the positive control: a purge stuck closed withholds every tenant and fails four of the eight hold tests, which is the only thing that separates a working control from commendable caution. One pre-existing test (`test_storage_tiering.py`) pinned the 90-day tiering DELETE and now reads the shared constant; `test_retention_worker.py`'s session double could not answer the new holds query, so the worker failed closed and five of its cases failed — the double was taught the query rather than the fail-closed behaviour relaxed. |