Security Audit Operations
Last Updated: 2026-11-12 | For: Production administrators and security reviewers | Level: Advanced
The configured workflow audits confidentiality and account-compromise risks at https://mofacts.optimallearning.org. Its configuration does not establish that a scheduled run completed successfully; review the stored report and its freshness. A red report is evidence for remediation. The audit does not stop services, change the reverse proxy or UFW, block traffic, deploy code, or affect real accounts. Authentication probes may change only their dedicated synthetic fixtures.
The scanners are separate from application deployment, but their sanitized reports are delivered to MoFaCTS through one authenticated ingestion endpoint. Administrators review them at /admin/security-audits. The audit still has no authority to deploy, reconfigure, block, or remediate production.
V2 candidate status (2026-11-11)
The application source now activates the V2 receiver and V2 admin presentation
at the existing admin URL. It accepts only /internal/security-audits/v2,
keeps the existing admin guards and method names, and uses separate private V2
storage. Existing V1 records are preserved without a current application reader.
The exact 1c72734b image is now accepted on staging2; production deployment remains excluded.
Production still runs its V1 receiver. The scheduled production workflow stays
on reviewed scanner commit 09f0c32fc2af6d220c699dd805133df3c0507c2e
until a separately authorized coordinated production cutover. Its report source
identity is the scanner checkout, which can differ from the workflow commit.
See the scanner cutover procedure
and the current acceptance ledger.
The five unidentified publication selections were replaced with explicit progress-privacy cases under the operator's approval. Source-owned intent, allowed-user baselines and real-Mongo publication tests pass; a timeout, missing route or unrelated application error still cannot establish denial.
The exact Meteor-accepted candidate 1c72734b is built, image/source-verified and published to the separate staging2 repository. Its isolated candidate/previous/candidate trial passed all three generations' list/latest/get/download and exclusion checks, preserved both stored report generations, and restored identical candidate download hashes. Cleanup and unchanged deployed MoFaCTS/JEDM services were independently verified. P1-16 is closed. The live staging2 V2 application/receipt/host cutover and a separate live V1 rollback/V2 restoration also passed. Independent stored-record hashes and restored JSON/HTML download hashes match; Mongo, Redis and unrelated JEDM services were unchanged. The actual ingested exposure report contains 22 PASS and one honest UDP evidence-incomplete ERROR: all seven external UDP results were inconclusive, not proven closed. Production is unchanged. See the linked ledger for exact image and evidence hashes.
The V2 host/access replacement carries frozen raw measurements plus the shared V2 receipt through the same restricted SSH command. Its actual restricted-channel acquisition and receipt MATCH passed with all eleven host controls PASS, plus arbitrary-command and PTY denial. Staging2 now selects the V2 checker at c79735ff; its SSH/sudo/Tailscale permissions are unchanged. The corrected live selected-authorization qualification passes 29/29 cases, and both disposable fixture attempts have independently verified account/role/data cleanup. The prepared scope is four disposable staging2 accounts, only two teacher grants, reuse of the existing administrator and exact cleanup; the operator explicitly approved this ordinary synthetic qualification scope on 2026-11-11, so further permission is not needed for those fixtures. See the fixture review. The bounded access delegation approved 2026-11-12 now authorizes necessary minimum staging2 audit permissions under Section 0 conditions, including the prepared report-signing binding and separate Desktop-held encryption key. Purpose, target, custody, lifetime, owner, verification and revocation must be documented; eligible changes need no additional approval. The initial GitHub read-only boundary no longer blocks that scoped report writer. The manual V2 workflow is now published, with the single staging2 signing binding and a separate operator-held RSA-3072 pair installed. Saved-key round trips, binding withdrawal and actual HTTPS signer checks pass (correct signer reaches invalid-report validation; wrong signer is denied); no report was accepted by those probes. GitHub currently cannot start jobs because of its billing lock, so fresh hosted audit acceptance remains unexecuted and no paid change is proposed. P1-12 now closes: ten superseded current V1 scanner files and duplicate builders are removed; historical defect witnesses use immutable test-only Git source, while production keeps its reviewed historical checkout. Phase 1 has 16 of 17 deliverables complete; its production identity/cutover gate remains unexecuted. The operator selected standard verified release recording on 2026-11-12. The canonical production command sheet now checks project-local pinned Node and SSH trust before deployment, then calls the verified record/history/receipt owner after its existing participant checks. Only those integration hunks are published; unrelated command-sheet edits are preserved. Node 24.22.1 was installed from the official checksum-verified, validly signed Windows archive without changing global PATH. The actual local preflight, existing-record format, seven focused Windows cases and combined Linux audit/source gate pass (843 passes, zero failures, 18 conditional skips). No production command, deployment, authority write or new Meteor invocation has run; live acceptance still needs production authorization. See the operating baseline and report-writer review.
Operator execution without GitHub billing (2026-11-12)
GitHub billing is unavailable. The four repository Security checks now run through
the standard operator Linux executor; the optional GitHub Security workflow is
manual-only. Source dff28386 passed full-history gitleaks (869 commits), both
dependency policies and source/security contracts (474 passes, nine conditional
skips, zero failures). This is operator evidence, not a hosted check or release
qualification. See the repository executor.
The operator staging2 exposure sequence
now composes the same qualified V2 owners without GitHub scheduling. Actual source
750523d9 obtained the existing restricted host command/MATCH receipt, assembled
and encrypted a new exposure report, submitted it once (HTTP 201), and verified
admin list/get plus JSON/HTML downloads. Saved-key decryption reproduced the exact
JSON download bytes. Owned plaintext was removed after verification; encrypted
recovery and content-free receipts remain in private operator evidence.
That report has 18 PASS, zero FAIL and five ERROR: the known incomplete UDP
inventory plus four HTTP checks that timed out after Linux DNS became unavailable.
Bounded OS/Node probes confirm EAI_AGAIN/ETIMEOUT; Windows signer/browser readback
worked. The final assertion correctly exits 1. This proves workflow integrity,
not complete security clearance. No DNS setting, production access, new grant,
application deployment or Meteor invocation occurred; six Meteor slots remain.
Phil subsequently approved restarting only Ubuntu-MoFaCTS. Its normal DNS
resolution recovered without global WSL/Docker restart or DNS configuration
changes. Docker engine and MoFaCTS/Mongo/JEDM API container lifecycle metadata
stayed unchanged; the scheduler continued its pre-existing restart loop. The
four affected HTTP controls now PASS through the same shared scanner probes
and evaluator, without repeating port/TLS/host/authorization checks or rewriting
the earlier report. Recovery source is a0fab958; 25 focused Windows cases,
full lint and Linux V2 source tests (309 passes, six conditional skips, zero
failures) pass. The known UDP gap remains. Six Meteor slots remain; P1-04 and
the production exit gate still require explicit production authorization.
Read-only production observation (2026-11-12)
The initial helper needed Windows ProgramData in its restricted SSH environment;
that source correction is tested without inheriting credential/provider values.
The initial two-minute image read stayed incomplete. Phil then approved one
explicit ten-minute diagnostic, and committed observer source a8cc2559 completed
it: 379,755,008 bytes in 310,803 ms, three successful metadata reads, process exit
0, zero stderr and confirmed termination. The final read verified the same running
container. No image layers were retained.
Observed source 18677291993b66b890a090e55d65f703b45ff653, Linux/amd64 and config
sha256:214793c9cf7534eb4568934fc83843328b202a23b99d108fc323ead2ceafb697
match the already approved November 5 P1-D2 baseline exactly. The older bddd0895
local authority remains unchanged; its mismatch is already reviewed. Do not ask
to choose that rollback target again or invent missing build/acceptance history.
No production service/configuration/data, app request, deployment or release
record write occurred.
Source 2343b03d records acceptance and explicitly selects the measured ten-minute
remote budget in the standard post-acceptance recorder; shared/local and bare
observer defaults remain two minutes. Nine focused Windows cases, full lint and
Linux V2 contracts (313 passes, seven conditional skips, zero failures) pass.
No Meteor invocation was consumed; six slots remain.
Phase 1 stays 16/17 because actual post-deployment production recording/cutover has not run. P1-D9 was approved on 2026-11-12: continue Phase 2-4 source/staging2 work after their applicable checks, and combine per-phase live production gates into one final coordinated cutover. Every acceptance criterion and separate final production authorization remains required. See the approved sequence and observation acceptance.
Phase 2 source comparison milestone (2026-11-12)
Source c99de5d9 adds source-owned login/reset comparison intent and sanitized,
versioned response-equality/timing proofs without changing the active catalog or
stored report meanings. Typecheck/lint and Linux source gate pass (321 passes,
seven conditional skips). One authorized isolated Meteor run passes: 1,021 server
and 1,249 client tests, zero failures; five standing invocations remain. Both new
registered shared-contract cases passed. Source/isolation/limits/cleanup and local
service continuity are verified. This is source/runtime evidence, not a deployment
or complete live reset/expiry acceptance. On 2026-11-12 the operator approved
reusing the connected inbox. Existing staging aliases route there; a bounded
one-account inspection received a fresh reset email with the exact staging2
link/recipient. The first unescaped-address lookup failed before reset delivery;
the same fixture was recovered, its guarded deletion and absence were verified,
and its DPAPI credential was removed. Only the selected reset body was read for
acceptance; no reset link was followed or password changed. No production
credential recovery, mailbox setting change or new provider permission occurred.
The private message-selection/link validator has four new Node cases; full lint
and Linux source checks pass (325 passes, seven conditional skips). Five Meteor
runs remain. Complete protocol leases, unused aged tokens, replay/session checks
and unattended transport integration are still pending. See the
comparison procedure.
Phase 2 reset and logout continuation (2026-11-12 local)
P2-06 publication isolation and P2-07 download mint/content/replay are checked off in the written plan using their accepted selected-case and real-Mongo evidence. Future Phase 3 registry scenarios do not reopen these Phase 2 proofs.
Reset/session source f3d21e74 added six reset cases and a versioned positive,
at-most-30-day lifetime proof. Its native run failed one new fixture because a
mixed-case seeded email did not match canonical identity. The corrected source
efa3bd73 passed 1,024 server/1,249 client tests, zero failures/hooks, 15/seven
existing pending. Actual native password checks confirm unused-expired rejection,
current reset, replay denial and stored-session-token removal. The isolated
synthetic age is not live wall-clock/mailbox-expiry or transport-revocation proof.
Source/isolation/resources/cleanup and unchanged local services were verified.
Three standing Meteor invocations remain; the failure consumed its own slot.
Operator adapter 524aec55 passed the live staging2 logout subcase against the
unchanged 1c72734b app: original and same-token clone baselines worked, logout
completed, cloned guarded-method access and fresh same-token resume were denied,
and a new password login worked. Job-owned sessions/contexts were logged out and
closed. No password, role, content, provider or production change occurred.
Full lint and Linux source gate pass: 335 passes, seven conditional skips.
Complete live reset/expiry, password-reset session revocation, protocol leases,
coverage and catalog/receiver integration remain open.
The original leakage implementation already distinguished normal account storage from unintended disclosure checks. A literal reading of its broad title was withdrawn; no finding establishes that Daybreak was wrong or authorizes weakening the check. On 2026-11-13 Phil rejected the IMAP approach and directed removal of the optional reader. Its package, lockfile, tests, local installed dependencies, Docker exclusion and supplemental test inputs are removed. Use the existing connected Gmail access for exact synthetic staging2 recipients, with bounded metadata selection and a private transient capability handoff. No new mailbox password, provider grant or production/recovery read is needed. See the connected-mail procedure.
The removal passes full lint (330 surfaces) and Linux source-security (335 passes, zero failures, seven existing skips). No Meteor run was used; three slots remain. Operator-assisted Gmail observations can qualify live scenarios but do not prove unattended acquisition by the standalone scanner. Keep that limitation visible while continuing the remaining reset/session, lease and coverage work. Historical IMAP qualification is superseded and is not live-mail acceptance.
Connected reset/session acceptance (2026-11-13)
The existing connected Gmail access supplied one exact synthetic reset message. The live staging2 scenario passed all three current-reset cases, including replay rejection. The new password worked and the old password was denied. Two independent pre-reset sessions had distinct capabilities and successful protected baselines; both lost guarded-method access and both old capabilities failed fresh resume. A fresh identity-bound session had 29.999981550925927 days remaining, within the shared positive/30-day bounds. P2-10 is checked off in the sole worklist.
Both disposable accounts were removed. A separate fresh admin session verified their absence; private body/capability handoffs were consumed and removed. No new provider credential, production access or application deployment occurred. The initial one-session observation is retained as narrower evidence, not the complete two-session proof. Genuinely aged unused-token expiry remains open.
The connected adapter uses one exact Gmail X-Received ingestion timestamp; missing/malformed/duplicate metadata fails before body reading. The observed connector internal_date had only whole-second precision; the request window is not rounded or moved. Full lint passes and the Linux source gate passes 349 cases, with zero failures and seven existing skips. No Meteor run was consumed; three remain. Shared environment lease/recovery tests pass; integration across all protocol families and the catalog-complete authentication artifact remain pending.
Protocol source milestone and current decision (2026-11-13)
Committed source fe08499c passes the full isolated Linux Meteor suite: 1,027 server and 1,249 client tests, zero failures/hooks, 15/seven existing pending. All three registered passwordless/catalog cases pass individually. Clean source, measured 6-GiB/two-CPU isolation, unchanged original local services and owned Docker-resource cleanup were verified; all 26 native proof files were retained and byte-verified. This is native/source evidence, not GitHub or image acceptance. Two of the ten authorized Meteor invocations remain.
P2-01 is checked off: source intent references all 22 authentication controls and 68 unique verification entries without changing the accepted 29 selected cases. P2-06/07/10 stay closed. The active producer still truthfully leaves unintegrated protocols ERROR. Live aged expiry, remaining acquisitions and the complete artifact are not closed by source coverage.
The subsequent Node-only protocol binding/process work passes five focused Windows cases, full lint and the Linux source gate (364 passes, zero failures, seven existing skips). Seven scoped credentials can be prepared for eight actor bindings including the existing operator; no new live account is created by the parser. Each child gets only its necessary subset. The shared environment owner preserves Linux's allowlist and explicitly supports Windows browser bootstrap paths; provider/signing/fixture environment data and Node injection are rejected.
P2-D2 is resolved by Phil's laboratory requirement clarification: learning progress intentionally grants return access in temporary passwordless experiments. Participants must not need remembered passwords or personal identifying information. The stricter marker-gate recommendation and inferred defect are withdrawn. No application gate, account data or migration was changed. Audit the intended participant/progress/experiment boundaries while preserving that allowed return path; missing markers alone do not establish a failure.
The repository already has Prolific-specific start/resume methods, a participation service and a server-side API client. Phil identified API validation for genuine Prolific entry as a possible improvement. Review those existing owners separately; this does not replace general laboratory participation, authorize provider actions or block the current audit. Accepted P2-01/06/07/10 and slot 8 remain closed. Two Meteor invocations remain; this documentation correction consumes none.
What runs
- Configured production daily schedule at 06:00 UTC: external network and internal host exposure controls; a schedule is not proof of execution while GitHub is locked.
- Monday at 06:00 UTC: all four sections, adding authentication and repository/image controls.
- Manual Actions runs:
exposureorfull.
The external runner resolves every A/AAAA address, scans public TCP and selected UDP ports, validates redirects, certificates, TLS protocols/ciphers, WebSockets, and security headers. An exact UDP open result fails. Silent open|filtered results pass only when the protected host evidence simultaneously confirms zero unexpected UDP listeners, zero Docker UDP publications, active default-deny inbound filtering, and zero unapproved UDP allow rules; unsafe or missing host evidence becomes FAIL or ERROR. TLS review evaluates enumerated cipher entries rather than matching normal compressors: NULL output. The forced host command checks sockets, Docker publications, UFW, the active Apache HTTPS virtual host and upstreams, MongoDB, Redis, the running image digest, and Sidecar bindings, reporting stable sub-probe outcomes and unexpected listener endpoints. Its socket policy permits the tailscaled process on wildcard UDP 41641 and its TCP peer listeners on Tailscale's reserved address ranges. It also recognizes a concrete-interface DHCP client on UDP 68 owned by systemd-networkd. Other unexpected listeners remain findings. Full runs add deterministic Playwright/DDP tests with synthetic accounts and separate runtime/development dependency checks. Development-only npm advisories remain visible as maintenance evidence but fail production posture only when an exact package and advisory ID has a reviewed build/CI exposure in the checked-in policy; runtime dependency and image findings remain independent. Authorization failures use stable semantic probe IDs, and passwordless containment reports each sealed-target boundary with impact-specific severity; anonymous resume-token issuance is expected to pass.
The report's source revision identifies the audit workflow checkout. It does not claim that production was deployed from a clean working tree.
The Sidecar is not part of scheduled execution. It remains available only for manual investigation with synthetic credentials, with ports 8931 and 8932 bound to 127.0.0.1.
Reports
Every run creates a bounded, sanitized canonical JSON report. The runner stores it in MoFaCTS and also encrypts it with RSA-OAEP-SHA256 and AES-256-GCM as a 90-day recovery artifact. Plaintext and raw scanner output are deleted from the runner.
- A
FAILis a security finding. It does not make production unavailable and does not fail the workflow. - An
ERRORmeans the result is not authoritative. Missing tools, malformed output, and unavailable fixtures enterexecutionErrorsand fail the workflow after report storage. A completed but inherently inconclusive probe stays visible asERRORwithout being counted as a vulnerability or workflow execution failure. - Failure to store the report makes the workflow fail.
- Workflow logs contain only generic messages and encrypted-artifact metadata.
The admin page receives only report summaries. Full findings are available through five-minute, single-use JSON and escaped HTML downloads. Reports expire automatically after 90 days. Exposure reports show a stale warning after 36 hours; full reports show one after eight days.
Download the artifact from its Actions run. From C:\dev\MoFaCTS\mofacts, decrypt it into a new file with the operator-held private key:
node scripts/security-audit/decrypt-report.mjs C:\path\report.encrypted.json C:\path\report.json C:\secure\mofacts-security-audit-private.pem
The decryptor authenticates the ciphertext and verifies the report's canonical SHA-256. Delete plaintext copies after review.
Protected setup
The protected GitHub environment production-security-audit contains the production Tailscale MagicDNS host, dedicated audit SSH user/key and known-hosts line, limited Tailscale OAuth client ID/secret for tag:ci, synthetic-account fixture JSON, dedicated IMAPS password, and AUDIT_REPORT_INGEST_SECRET. The production application receives the same value as MOFACTS_SECURITY_AUDIT_INGEST_SECRET. AUDIT_REPORT_ENCRYPTION_PUBLIC_KEY is a protected environment variable containing the RSA public key; it is not secret.
The matching private key stays only with authorized operators. It must not be stored in GitHub, MoFaCTS settings, the production host, or source control. Generate a 3072-bit pair with:
openssl genpkey -algorithm RSA -pkeyopt rsa_keygen_bits:3072 -out mofacts-security-audit-private.pem
openssl pkey -in mofacts-security-audit-private.pem -pubout -out mofacts-security-audit-public.pem
openssl rand -base64 48
On the host, install deploy/security-audit/host-exposure-audit.sh as a root-owned forced command, install host-listener-policy.awk and host-firewall-policy.awk under /usr/local/libexec/mofacts-security-audit/ with root:root ownership and mode 0644, and install security-audit.conf.example as /etc/mofacts/security-audit.conf with root:root ownership and mode 0600. Reinstall the host command and both policy files after scanner updates. Configure tailscale0 as the SSH management interface, the IPv4 and IPv6 tailnet ranges as the management CIDRs, and current service/container identities. The firewall policy requires active UFW, default-deny inbound, SSH only on that private interface from those ranges, and public TCP 80/443 only. Before enabling UFW, retain a public maintenance session and prove both operator SSH and the forced audit key through the production MagicDNS name. The production overlay sets HTTP_FORWARDED_COUNT=1 for the single trusted Apache proxy so Meteor's IP throttles use the originating client rather than Docker's private proxy address. Production disables Meteor inline runtime scripts and enforces the tracked CSP; only reviewed dynamic style attributes retain the narrow style-src-attr 'unsafe-inline' exception, and new external instructional image, media, or frame origins require explicit CSP review. MOFACTS_REDIS_PASSWORD must be the same URL-safe private value in the Compose environment and the root-only audit configuration; the audit verifies authenticated success, unauthenticated denial, and no Redis host publication without printing the value. The SSH key must permit no shell, PTY, forwarding, agent forwarding, X11 forwarding, or arguments.
The production CSP permits WebAssembly compilation for Anki import through the wasm-unsafe-eval source in script-src; the separate unsafe-eval permission remains prohibited. Voice previews use the direct https://docs.cloud.google.com/text-to-speech/docs/audio/<voice>.wav documentation sample for the ten offered voices. Permit https://docs.cloud.google.com in media-src; the preview no longer depends on a redirect through cloud.google.com. Existing additional media allowances require their own review before removal.
On October 9, 2026, staging2 candidate a3155584 passed synthetic upload ownership/transport probes, normal package/media upload, Chrome sample playback and inline failure display, visible theme confirmation, learner reload/resume, Anki analysis, hosted fonts and representative YouTube checkpoint playback. CSP continued to deny ordinary evaluation and inline scripts; login/reload/reconnect did not recreate x_mtok. The upload/voice/theme repair batch is accepted for this scope. This does not close Phase 0, qualify production or establish complete provider/surface coverage. See the repository durable audit status for image identity, verification limits and remaining work.
The canonical policy is C:\dev\mofacts_config\production\apache\000-default-le-ssl.conf. YouTube lessons use Plyr and require https://www.youtube.com/iframe_api plus https://www.youtube.com/s/player/ in script-src, https://i.ytimg.com/vi/ in img-src, and https://noembed.com/embed plus the control-icon file https://cdn.plyr.io/3.8.4/plyr.svg in connect-src. Review the versioned icon URL when upgrading Plyr. Keep playback frames on the existing https://www.youtube-nocookie.com allowance. These are explicit HTTPS path permissions; inline scripts, general JavaScript evaluation, and wildcard sources remain prohibited. After an authorized deployment, check the served CSP and confirm that video lessons initialize, play, present timestamp questions, and enforce their completion gate.
Before valid host JSON exists, transport failures are reported with one bounded category: tailnet-connection-failed, ssh-identity-configuration-failed, ssh-transport-failed, forced-command-rejected, or host-output-invalid. They produce complete internal ERROR controls without retaining raw SSH diagnostics. For tailnet-connection-failed, verify the protected environment contains TS_OAUTH_CLIENT_ID and TS_OAUTH_SECRET for an OAuth client allowed to create ephemeral tag:ci devices.
Production builds carry the exact Git commit in the OCI org.opencontainers.image.revision label. The authoritative production deployment verifies that label before recording the release, while the host audit reports the independently observed running-image SHA-256 identity.
The opt-in build-source capture prototype has a recorded real source capture, Windows-to-Linux round trip and structural image smoke. Full application smoke, source-proof binding and qualification gates remain open. It is not installed in the production command sheet, does not issue release assurance, and ordinary builds do not require its metadata.
The separately authorized collect-tests action verifies the captured source and creates an owned derivative containing reviewed supplemental test inputs. Missing inputs, unsafe paths, links, conflicting build bytes or modes, and acquisition failures stop collection; an existing derivative is never replaced. Success is test-inputs-captured-unqualified. Collection starts no services and runs no tests or image build. The October 9 source checkpoint reports synthetic tests; real collection on the supported Linux lane, complete runner/evidence checks, stronger sealing and final image source proof remain open.
The complete setup and configuration inventory are in the repository's periodic audit deployment guide.
Synthetic fixtures
Normal account sessions have a configured 30-day maximum lifetime and use session storage. Authentication errors deliberately avoid confirming whether an account exists. The audit checks session expiry, revocation and account isolation with synthetic identities; a successful probe does not justify disclosing account details in logs.
Use a dedicated empty tenant containing two learners, two teachers, a synthetic audit administrator, reset, expiry, and lockout identities plus an existing incomplete passwordless-study participant. Provide only minimum synthetic course, history, dashboard, experiment, settings, route, and export state. Add recognizable canaries to synthetic records so cross-user leakage can be detected without retaining raw payloads.
Use a dedicated IMAPS mailbox for reset links. Retain messages older than the configured one-hour lifetime so expiration can be tested. The scanner never treats a merely prior-run message as expired without checking its timestamp. Failed authorization probes appear under deterministic semantic IDs; reset-token and throttle failures use fixed subprobe IDs. Evidence is bounded and never includes fixture arguments, identities, tokens, or response bodies. Brute-force probes run last and may affect only the lockout canary and unique synthetic throttle identifiers.
First run
Run the first manual full audit only after the forced host command, UFW management CIDRs, encryption public key, matching ingestion key, mailbox, and fixtures are installed. Passwordless experiment sessions are expected to receive anonymous resume tokens; the audit tests that they remain contained to the sealed experiment target. Other initial findings may include unauthenticated Redis or missing CSP.
Do not change those systems while interpreting the report. Open separately scoped remediation work, verify it, and let the next scheduled report record the result.
Staging2 audit connection qualification (2026-11-15)
The separate audit connection is installed on staging2 from committed source
0592752b. Nine exact-image cases pass, including actual service startup and
shutdown. Live HTTPS metadata, unauthenticated denial, private socket ownership,
a real systemd restart and unchanged existing application/JEDM containers passed.
The connection container has networking disabled; Apache reaches a separate HTTP
Unix socket. It cannot access the private scanner socket or connection state.
The service has no Gmail, database, shell, signing or report-decryption credential.
The private connection package is registered; the operator's one-time consent is pending. Installation is not proof of cloud tool access, real reset-message carriage, unattended scheduling or a complete authentication audit. Those gates remain open. Existing progress-backed passwordless return remains intentional and unchanged. No production or original-staging change is included.
The connection procedure owns access, recovery and revocation. The sole audit checklist records exact evidence and remaining acceptance.