red_lead/blue_lead never saw a plain "All Tests" list — they were always
routed into their own review two-queue view or "My Tasks", with no way to
see every test and who's actually working it. Adds an always-visible
section at the end of the Tests page for leads showing every test with
both Red and Blue assignee columns, so they can monitor the queues and
spot a stuck ticket or an overloaded operator.
The list endpoint (GET /tests) now joinedloads the assignee relationships
and resolves usernames the same way the detail endpoint already does, so
the new section doesn't depend on GET /users/operators (lead/manager-only).
- BT Start Date had the same bug RT Start Date did: pushed on every round,
clobbering round 1's real pickup date. Now only pushed on round 1,
mirroring the "first round survives every reopen" rule.
- push_test_event's assignment branches (red_executing, red/blue reviewer,
blue_evaluating reassignment) only read the cached jira_account_id and
silently no-op'd if it was never populated — which is exactly what
happened to a blue_lead who hadn't logged in since Jira was configured,
making it look like Blue's reviewer reassignment was broken when Red's
wasn't purely because that particular red_lead happened to have logged
in already. Extracted the live-lookup fallback that only
push_assignee_update had into a shared _resolve_jira_account_id(), used
by every assignment call site now.
- Tests created without a campaign parent ticket now get a "standalone-test"
label, so they're identifiable in Jira without cross-referencing Aegis.
- The comment posted on every red/blue submission was a generic one-liner
with no data — only a reopened round got a full data dump (via
push_round_archived), so a round that was approved outright left no
record in Jira of what was actually done. Now every submission comment
includes the round number, procedure/tool/attack success (red) or
detection/containment result (blue), and summary.
- RT/BT Start/End Date are genuine Jira "datetime" custom fields (confirmed
via issue_editmeta), not plain dates. Pushing a bare "YYYY-MM-DD" string
made Jira default the time-of-day to midnight, so RT Start Date and RT
End Date ended up showing the same meaningless midnight timestamp instead
of the actual execution window. Now sends a full ISO datetime.
- The test detail page showed "RT: Unassigned" for the operator who WAS
correctly assigned, because GET /users/operators (the only source
AssigneeControl used to resolve an assignee ID into a username) is
restricted to leads/managers — a plain operator has no other way to
resolve their own ID. TestOut now resolves and includes the assignee's
username directly (red/blue tech + reviewer), so the badge no longer
depends on a permission-gated list the viewer might not have access to.
- execution_start_time/execution_end_time/detection_time/containment_time
were captured from a datetime-local input with no local->UTC conversion,
then stored/displayed as if already UTC — off by the browser's offset.
Frontend now converts properly in both directions; backend schemas
defensively normalize any tz-aware input to naive UTC as well.
- push_rt_submitted was pushing datetime.utcnow() (whenever the submit
button was clicked) into Jira's RT Start/End Date fields instead of the
operator-entered execution window. Now pushes the *first* round's
execution start (recovered from round_history across reopens) and the
*current* round's execution end. Removed push_rt_started, which only
ever pushed a meaningless click-timestamp.
- Reopening a test now adds a "reopened" Jira label (read-modify-write so
existing labels aren't clobbered), on top of the existing round-archived
comment and status push.
Only the per-round verdict/timing fields (attack_success, execution
start/end for red; detection_result, containment_result, detection/
containment time for blue) get cleared on reopen — the operator must
answer those fresh for each round, and the prior round's answer stays
visible via round_history + the archived Phase Timeline card. Free-text
fields (procedure_text, tool_used, red_summary, blue_summary) are no
longer cleared: the operator edits them in place for the new round
instead of retyping everything from scratch.
- Reopening a test for rework (red or blue) now archives the round that's
ending into a new test_round_history table before resetting the live
fields — procedure/results/detection data is preserved instead of
overwritten, and Phase Timeline renders every past round alongside the
current one, so Blue Team keeps visibility into earlier Red attempts.
- Each archived round also posts a permanent Jira comment (push_round_archived)
since Jira's custom fields only ever show the latest value once the next
round resubmits — the comment thread is what preserves full history there.
- push_pause_event no longer transitions the Jira issue to "On Hold" — a
timer pause is not a real Hold, and flipping Jira status for it was
misleading. Only the explicit Hold action (push_hold_event) does that now.
- reopen_red_review starts the timer paused (paused_at set) instead of
immediately running, since the operator hasn't resumed working yet —
blue already did this via blue_work_started_at, red needed the same
behavior via the existing pause/resume mechanism.
- update_test_red/update_test_blue now lock edits to the actual assigned
operator for leads too, not just techs — a lead could previously edit
another operator's in-progress test. Mirrored in TeamTabs.tsx.
- push_test_event reassigns the Jira ticket to the original operator
(not the reopening lead) when a test returns to red_executing or
blue_evaluating for rework, instead of leaving it on the lead.
The red_lead reviewer stayed assigned in Jira through blue_evaluating
(and the blue_lead reviewer through in_review) since push_test_event
never touched the assignee on those transitions — only red_executing
and the review gates did. Both are hand-offs with no single owner yet,
so the assignee is cleared instead of left stale.
Also fixes push_bt_work_started never assigning the blue_tech who
actually picks up the queued test — it only moved the status to
In Progress before.
- Fix the red/blue draft form losing unsaved fields (e.g. execution_start_time) whenever an evidence upload refetches the test — the hydration effect now runs once per test, not on every refetch.
- Add color-scheme: dark so native date/time picker popups match the app theme instead of rendering white.
- Fix _STATE_TO_JIRA_STATUS: real Jira statuses are 'To Do' and 'Red Team test review', not 'To-Do' and 'RT Test Review' — confirmed live against the actual workflow transitions, which is why Submit to Blue Team never moved the ticket past 'In Progress'.
- Tempo worklog sync was silently blocked by TEMPO_ENABLED defaulting to False with no admin-facing toggle, even after configuring a Tempo token via Settings. Now bypassed once an admin or personal token is actually configured, mirroring Jira's DB-backed enabled flag.
- Hold now actually pauses the running phase timer (paused_at), and Resume accumulates the held duration into red/blue_paused_seconds — previously the timer kept counting through a hold.
lookup_user_jira_account_id() only ever searched by Aegis login email,
ignoring the jira_email override field that exists precisely for users
whose corporate Atlassian email differs from their Aegis login email
(the rest of jira_service.py already documents and uses this priority
order). Found via jesus-huertas, whose real Jira account uses a
different email than his Aegis account.
Admin can no longer be picked as a red_tech/blue_tech assignee — it
administers the site, it doesn't operate tests. Applied to the
eligible-assignee list on both the picker UI and the assign endpoint's
validation, and dropped from the GET /users/operators pool.
push_assignee_update() now falls back to a fresh Jira account-id lookup
when the assignee hasn't logged in yet (so jira_account_id is still
empty), instead of silently no-op'ing — assignment now syncs to Jira
immediately regardless of whether the assignee has ever logged into
Aegis before.
Admin administers the site — it no longer has any override on
validate-red/blue, review-red/blue, resolve-dispute, request-discussion,
reopen, hold, resume, assign-operators, or classification-update.
Introduces require_any_role_strict(), a role dependency without the
global admin bypass, so these specific endpoints truly exclude admin
instead of only removing it from the (redundant) role tuple.
Managers gain the ability to assign red_tech/blue_tech operators
(POST /tests/{id}/assign, GET /users/operators) alongside leads, since
that's coordination, not resolving a ticket.
Also enlarges and repositions the operator-assignment controls next
to the Start Execution button, and fixes a literal '\u2014' rendering
as text instead of an em dash in the test detail technique line.
GET /system/jira-config was admin-only, so non-admin users hit a 403
and the frontend silently fell back to a hardcoded jira.atlassian.com
base URL when building /browse/{key} links. Opened it to any
authenticated user (still no secrets returned).
Also drops the redundant 'security-test' Jira label, and adds a
page-size selector (10/25/50/100) and a 'no existing test yet' filter
to the Test Catalog.
Adds GET /users/operators (leads+admin) and wires a lead-only assign
control into the test detail header, calling the existing but
previously unreachable POST /tests/{id}/assign endpoint. Manual
assignment now also pushes the Jira ticket assignee immediately
instead of waiting for the operator to start execution.
Also adds test.platform as a kebab-case label on Jira ticket creation.
Jira Cloud's REST API rejects a bare string for select-type custom
fields with 'Specify a valid id or name for <field>' — confirmed live
against the real Jira instance. This silently broke Severity, Data
Sensitivity, Attack success, Attack detected, and Attack contained on
every ticket create/update, masked until now by the non-fatal
exception handling around each call.
The auto-create-ticket call was sending customfield_11233 (a field that
doesn't exist in the project) instead of the real Data Sensitivity field
customfield_11814, which made Jira reject the WHOLE issue and silently
block ticket creation for every new test.
Also replaces Aegis's invented 4-tier data_classification scheme
(public_release/general_use/confidential/restricted) with the 6 values
Jira's Data Sensitivity field actually uses (public/general_use/
internal_use_only/trusted_people/customer_content/pii), since that is
the classification the organization actually applies. Includes a data
migration remapping existing rows and a defensive retry if Jira ever
rejects an unscreened custom field again.
atlassian-python-api takes limit, not maxResults. The TypeError was
silently swallowed at DEBUG level, so every non-admin user's Jira
account-id lookup failed even with a valid email and working admin
Jira connection.
The Coverage Matrix heatmap only ever showed the bare MITRE ID on each
cell (name only via hover tooltip), and colored cells from a continuous
score -> hex gradient with no discrete status concept — unlike the
Techniques page, which showed ID+name inline with a fixed 5-color status
palette. Unifying the two pages' visual style requires both pieces of
data to exist server-side.
- Adds "name" to all 4 layer builders (coverage/threat-actor/detection-
rules/campaign) — harmless extra field for Navigator exports, lets the
frontend show it inline.
- Adds "status" (the real TechniqueStatus value) to the coverage and
threat-actor layers specifically — the only two backed by an actual
TechniqueStatus concept; detection-rules/campaign scores don't
correspond to one and don't get this field.
GET /tests caps limit at 200, so any page relying on len(results) for a
"total" count (e.g. Validated Tests) silently under-reports once there
are more matches than the page size — the 200 shown was a page cap, not
the real total.
- Adds GET /tests/count, sharing its query-building with list_tests()
via a new _build_test_query() so the two can't drift apart.
- Adds technique_search (free-text on MITRE ID or name), attack_success,
detection_result, and validated_from/validated_to filters, so callers
aren't limited to state/technique_id/platform/creator.
Previously only red_executing (-> "In Progress") ever changed the Jira
issue's status; every other Aegis transition (red_review, blue_evaluating,
blue_review, in_review, validated, rejected, disputed) only posted a
comment, leaving the Jira board stuck on whatever status it started with.
Expands push_test_event() to a full TestState -> Jira-status table
covering the whole lifecycle (To-Do -> In Progress -> RT Test Review ->
Queued Blue Team -> In Progress -> Blue Team Test Review -> Validation ->
Done/Rejected/Dispute), matching the Purple Team Jira workflow. Adds two
new lifecycle hooks that the generic dispatch can't handle correctly:
- push_bt_work_started(): blue_evaluating covers both "queued, unclaimed"
and "actively being worked" — needs an explicit push when the blue tech
clicks Start Evaluation, or it never leaves "Queued Blue Team".
- disputed state push: a conflicting lead vote previously produced zero
Jira signal at all.
Also fills two real gaps: reopen_red_review and reopen_blue_review never
synced anything to Jira before. Their target status differs on purpose —
reopen_red_review pushes "In Progress" (Aegis resumes the same operator
immediately, no re-claim), while reopen_blue_review pushes "Queued Blue
Team" (Aegis clears blue_work_started_at, forcing a fresh "Start
Evaluation" click) — verified against the actual field-reset behavior in
test_entity.py rather than assumed symmetry between the two teams.
- start_execution now blocks a test from starting before its campaign's
start_date, closing a gap where the field was persisted but never
enforced
- GET /tests gains a reviewer_id "my reviews" filter for the red_review/
blue_review lead-gate stage, distinct from pending_validation_side
(which only ever covered the later in_review stage and ignored
assignment entirely)
- get_coverage_by_tactic now returns all 14 MITRE tactics in canonical
kill-chain order instead of an alphabetical, partial list — the
regular Dashboard and Executive Dashboard both consume this endpoint
and previously disagreed on order/completeness
- test-template listing now includes existing_test_count per technique
so the catalog can warn before creating a likely-duplicate test
Mirrors the existing Atomic/Caldera/Elastic/Sigma/LOLBAS import pattern —
D3FEND mapping import was the one source silently skipping the
review_required flag, so leads never got prompted to review techniques
that only gained new D3FEND coverage.
Maps Test.containment_result (contained/partially_contained/not_contained)
to Yes/Partial/No, mirroring the existing Attack Success/Attack Detected
field push in push_bt_submitted. Field ID was the last missing piece
blocking this from Block 4.
- /auth/refresh now allows a short grace window past expiry and checks
the blacklist, so an active session's silent refresh no longer fails
the instant its own token expires
- normal manager /approve flow now creates Jira tickets for the campaign
and its already-linked tests, matching the admin-only /activate path
- GenerateFromActorPayload now accepts start_date and threads it through
to the new campaign instead of silently discarding it
Now that only the manager sets start_date, and only at the moment of
approval (which immediately activates the campaign), a draft campaign
can never have a start_date. This made three things permanently
unreachable: the hourly _run_scheduled_campaign_activation cron job,
the activate_campaign 409/force future-date guard, and a query filter
hiding tests from scheduled-but-inactive draft campaigns. Removes all
three rather than leaving dead code behind. Also adds the missing
manager-cannot-directly-activate router test.
serialize_modification_request crashed toward a string 'None' instead of
JSON null for requests whose test_id was nulled by the SET NULL cascade.
Also corrects two comments that still described the debunked ordering
theory instead of the actual SET NULL fix, and fixes a test that
asserted post-cascade state without committing first (the ORM only
picks up out-of-band DB-side SET NULL on already-loaded objects after
expire_on_commit forces a re-read, matching real router behavior).
- Single admin Jira account stored in system_configs (jira.admin_email, jira.admin_api_token)
- Admin Tempo token in system_configs (tempo.admin_token)
- get_admin_jira_client() + has_admin_jira_configured() replace per-user auth in all lifecycle hooks
- lookup_user_jira_account_id() auto-discovers each user's Atlassian accountId by email on login
- auto_log_test_worklog() now uses admin Tempo token and logs both red+blue team time
- Settings > Jira: admin credentials fields + test buttons moved to admin Jira tab
- Profile section simplified: jira_account_id shown read-only (auto-detected)
Tar Slip (CWE-22) — 3 import services:
threat_actor, lolbas, caldera: add path validation before extractall()
to prevent malicious zip members with ../ escaping the target directory.
(sigma, elastic, atomic already had this protection)
Path Traversal (CWE-23) — professional_reports.py:
Add _assert_safe_report_path() check on all 5 report endpoints to
verify the generated filepath stays within REPORT_OUTPUT_DIR.
Open Redirect (CWE-601) — sso.py:
Validate IdP redirect URL scheme (must be http/https) before
issuing RedirectResponse, blocking javascript: and data: redirects.
DOM XSS (CWE-79) — 4 frontend pages:
Create src/utils/url.ts with safeUrl() that rejects non-http/https
protocols; apply to actor.mitre_url, ref.url, intel.url.
Sanitize framework name to alphanumeric-only before DOM insertion.
Restrict evidence MIME types to an explicit safe allowlist (png/jpg/gif/webp).
Hardcoded credentials (CWE-798):
verify_gaps.py, create_wiki.py: replace literal passwords with
environment variable reads (AEGIS_ADMIN_PASSWORD, GITEA_PASSWORD).
- config.py: move REPORT_OUTPUT_DIR from /tmp (world-writable) to /app/reports
to prevent CWE-377 symlink attack vector (B108, only real security issue)
- main.py: log startup seed failures instead of silently swallowing them (B110)
- Add # nosec annotations to intentional try/except patterns that are by design:
Jira integration errors, email failures, DetachedInstanceError, storage errors,
and Jira session timeout (all B110/B112 false positives)
- Add # nosec B105 to false positives where bandit misidentifies config key
names and masking strings as hardcoded passwords
- Add .bandit config to skip B311 in seed_demo.py (random used for fake
demo data generation, not cryptographic purposes)
Group 1 - Dual validation rejection (9 tests):
_check_dual_validation: any single rejection is a veto (r or b == rejected
-> rejected). Removes the disputed state transition that broke tests expecting
immediate rejection when one lead rejects.
Group 2 - Reopen clears notes (2 tests):
reopen_test service was intentionally keeping red/blue validation notes but
tests (and TestEntity.reopen domain method) expect them cleared. Align service
with domain entity behavior.
Group 3 - Audit integrity hash (2 tests):
log_action: call db.refresh(entry) after initial flush and before computing
the HMAC hash. Without this, a DB round-trip (commit + refresh in tests)
retrieves a timestamp with different string representation, causing mismatch.
Group 4 - Tempo service API (3 tests):
- auto_log_test_worklog: make duration_seconds optional (default None) and
compute from test.red_started_at -> updated_at when not supplied.
- Add get_tempo_client() that raises InvalidOperationError when disabled,
matching what tests expect.
- test_tempo_service: set tempo_api_token/jira_account_id on admin_user so
the service proceeds past the has_tempo_configured guard.
Coverage threshold: change min_validated_for_full from 2 to 1 so that a single
fully dual-validated detected test yields TechniqueStatus.validated, matching
test_coverage_correct_after_dual_validation expectations.
kitos