The /system/llm/ badge read llm_profile_health, which only the failover
Pool wrote. A task pinned to an explicit chain, a bound agent with
failover off, and the global provider all call the provider directly, so
a dead backend stayed "Healthy".
One policy now lives in llmpool.Registry.RecordCall, used by the Pool and
by the task runtimes: a success clears the state, a failover-eligible
failure trips it, context cancellation and deterministic 400s do not, and
a stream that already delivered output is left alone. The selection
carries the profile whose breaker the call updates, and a provider
already wrapped in the Pool reports nothing to avoid double counting.
The LLM recording page also explains an empty list while the llm_record
toggle is off, after the settings call resolves.
Originally reported in #257.
The outbox migration is 0006 because 0005 is already traffic rewrite rules. A shutdown cancel leaves an in-flight intent claim running, so the next start can reset it to open instead of recording the interrupt as done or blocked.
The task report and the findings export asserted every finding with no
qualification: nothing said how far exploitation actually reached (PoE level),
whether the claim was machine-verified (verification status and note), whether
counterevidence was recorded, or whether any traffic was bound. The web UI
badges all of it (VERIFICATION_META, the detail page), so the report — the
artifact a client actually reads — was the last surface still asserting bare
claims.
Render it from one place:
- report/quality.go (new): two constructors and two pure renderers, no I/O, no
DB handle, so the block is unit-tested without a database.
- Both surfaces call the same helpers, which is the point: the task report and
the per-finding export cannot drift apart the way duplicated formatting does.
- Missing is explicit: a finding node with no standalone row prints
"not assessed (no finding record)" instead of an invented zero — the honest
gap the existing coverage work (#207) also insists on.
- Labels match the shipped UI wording (Verification, Confidence rationale,
Counterevidence) and states are humanised ("Not assessed", "False positive"),
with an unrecognised value printed verbatim so a new state cannot silently
vanish from a report.
- Bound-traffic count appears once per document: as a Proof bullet in the task
report (it has no traffic section), and in the export only when nothing is
bound — otherwise the traffic section already states it.
- The detailed report is capped at 8 KiB in the task report only, rune-safe, with
the cap derived into the truncation message; the export is the canonical
per-finding document and is never silently shortened.
- The export keeps its document structure: the single-finding document gets
"## Detailed Report" (a section like its siblings), the multi-finding summary
gets a bold label.
Read paths only: no change to the write side of poe_level or verification_status,
no schema change (all five columns already exist), no API/DTO change (the widened
FindingMeta stays internal to the report path, and the findings list keeps
selecting only what it renders). One extra task-scoped query on GET /api/report,
because findings bodies stay off the list path.
docs/vulnerability-traffic-evidence.md §Export describes the new Markdown.
GET, POST, PUT and DELETE /api/traffic/rewrite-rules and
POST /api/traffic/rewrite-rules/{id}/toggle manage the rules behind
requireAuth. A rule is validated on save (400), created disabled, and
validated again before the toggle enables it. The value of a sensitive action
is never returned; an update sent without it keeps the stored value when the
stored rule has a sensitive action of the same type and header name at the
same position.
The manager compiles the enabled rules after traffic.Open and after every
change, and publishes the snapshot with Traffic.SetRewriteRules, an atomic
swap. An enabled rule the validation refuses is left out and reported as
load_error.
The proxy stores the snapshot and does not apply it: a test sends the same
request with and without matching rules and compares what the target
receives, what the client receives and what is recorded.
The Host refusal now explains that go-mitmproxy rebuilds the upstream request from the URL and drops the Host override. ReplayResult gains Recorded (false for an https passthrough host) and the sorted response Headers, both exposed by the route. Headers named in Connection and Proxy-Authorization are dropped. A body over the limit returns ErrReplayTooLarge, mapped to 413.
Traffic.Replay sends an edited request (method, URL, headers, body) through the local proxy, which records the exchange like any other, and returns the response bounded to 2 MiB. POST /api/traffic/replay exposes it to operators only (no agent tool). Invalid input is rejected with ErrReplayInvalid; hop-by-hop headers are dropped, redirects are not followed, and a Host that differs from the URL is refused because the proxy routes by Host.
The recording proxy now stores WebSocket and Server-Sent Events streams: one exchange per stream (101 websocket, or the real text/event-stream response) and its messages in stream_messages, with per-message and per-exchange limits and blob spill. Messages are readable through Traffic.StreamMessages, GET /api/traffic/stream and the traffic_messages agent tool; deletion and blob GC cover them.
getTask aggregated task_relations for the whole database on every 5-second detail poll to read one entry. TaskArchiveBlocker scopes the same predicate (shared with TaskArchiveBlockers through archiveBlockerScope) to one source task; listTasks keeps the global map.
Mechanical: gofmt -w on the 11 files gofmt -l listed on main. Besides spacing and alignment, gofmt sorted the imports inside one group in two test files and removed a trailing blank line in a third. The go job now fails, listing the files, when gofmt -l reports any tracked Go file.
getTask computed the list metrics of every task to read one; it now calls TaskListMetrics(expID), which runs the same laterals as TaskListMetricsAll with a WHERE on the exploration. listTasks keeps the global query. The latest_activity lateral ordered by its EXTRACT(EPOCH) output alias, so PostgreSQL read and sorted every activity row of each task; it now orders by activity.created_at and idx_act_latest serves the LIMIT 1. Both queries share one SQL constant.
The egress governor is now a current governor that SetProfile swaps: new requests use the new one, requests in flight finish on the old one, and one package-level ticker logs the counters of the current governor. WebFetch, the http custom tools and the enrich probe resolve it per request, and the agent prompt block reads the current profile. The noise_profile setting (off, quiet, balanced, aggressive) is applied at startup and wins over ARTEX_NOISE_PROFILE, which stays the fallback; PUT /api/settings validates, stores and applies it without rebuilding agents, and GET /api/settings/noise returns the per-host counters without resetting them. System settings gets a Target noise card.
ScoreRisk scores a pending tool call from fixed, additive signals (no model): shell destructive, privilege, secret, exfiltration, reverse shell, scanner and file-write patterns; file writes with a sensitive-path bonus; web reads with a mutating-method bonus; MCP, unknown tools and unreadable input never score low. The score is computed when a pending row is read and returned as risk next to the unchanged row fields. POST /api/intercept/pending/batch-decide decides up to 100 ids, each through Interceptor.Decide with the same approver and address as a single decision, and reports a result per id. The approvals page gets row selection, a confirmation that counts high-risk rows, per-id errors, and a risk badge with its reasons.
PromptOverride, WrapupOverride, WrapupMaxTurnsOverride,
WrapupTaskTimeoutOverride and WrapupTaskTimeoutTurnsOverride are wired
in server.New while restored tasks may already render prompts and build
settlement configs on other goroutines. go test -race ./server reported
races between server.New (server.go:195/207/215) and agent.renderSystem /
resolveWrapup / resolveWrapupTurns.
Same treatment as the tool hooks: unexported variables behind a
sync.RWMutex (one for the prompt hook, one shared by the four wrap-up
hooks), Set* setters and getters, and readers that copy the function
value under the lock and call it outside. A nil hook still means the
built-in default.
TestAgentHooksConcurrentSwap now covers all eight hooks.
Phase 2 of secrets at rest (#213), same AES-256-GCM scheme and key.
- llm_profiles.proxy joins the secret columns (AAD llm_profiles.proxy);
scanProfile decrypts it for both column lists, SaveProfile encrypts
it, an empty proxy stays NULL
- brave_search_api_key, tavily_search_api_key, web_search_proxy and
global_proxy are stored encrypted (AAD settings.<key>) through
GetSetting, SetSetting and InsertSettingIfAbsent; other settings stay
in clear
- the startup backfill, the no-key refusal and the wrong-key check
cover the new values; legacy plaintext is read as-is until backfill
API: the proxies are returned masked as a whole (placeholder plus a
*_set flag) by GET /api/settings, GET /api/llm/profiles and GET /api/llm.
On write the placeholder keeps the stored value, an explicit value
replaces it and "" clears it; the placeholder is never stored. The test
and model-list endpoints resolve the placeholder to the stored proxy.
POST /api/intercept/pending/{id}/always-allow derives an argv allow rule from the stored pending row (never from the body), at priority -1000, reusing an identical enabled rule, and leaves the request pending. A code-owned denylist of interpreters, launchers, package managers, orchestration and destructive programs refuses such rules at creation, in the manual rule form, and neutralises argv allow rules at matching; argv deny and ask rules still match.
Normalise and length-check ledger fields in db.RecordCoverage, frame the planner gap lines in the data envelope, cap the wrap-up (25) and report (200) gap lists, cap current rows at 2000 per task. Replace the activity budget marker with tasks.end_reason, written with the guarded terminal status and cleared on revive. Generic 500 message in taskCoverage; early return on an empty ledger in the planner block.
Workers record per (surface, risk_area) outcomes with the new record_coverage
tool (reported, no_issue_found, ruled_out, not_applicable, needs_follow_up).
Gaps are needs_follow_up entries, entries missing their required evidence or
reason, and goal vulnclass values with no entry (unrecorded_risk_class), so an
absent finding never reads as a clean target. Completeness marks timeout,
budget wrap-up, pause, abort and running tasks as incomplete with a caveat.
Surfaced in the planner context (one line per round, full gap list on the
timeout/budget wrap-up round), GET /api/tasks/{id}/coverage, the coverage card
and the Markdown report. Follows coverage_enabled; no new env var.
ARTEX_UPDATE_CHECK=1 lets the backend check the public GitHub releases API
for the latest tag at most once per 24 h (in-memory cache; failures are not
retried in the window) and report it in GET /api/health when it is newer
than the build version. The web UI shows a dismissible strip linking to the
Releases page. Unset (the default), no update request is made.
spawn_task now reads the target-gate scope of the calling session from the
tool context. When the caller is gated, the child task is created with the
gate on (D23) and with a copy of the calling task's block rules (D24). Pins
and allow rules are never copied, parent_ref is not consulted, and spawn_task
gets no new parameter. A gate-off caller, or a call with no scope, creates the
child exactly as before.
A conversation started by a trigger of a gated task is created with the gate
on (D25). A merged run is gated if any of its fires came from a gated task.
Pins are never copied. Fires of a gate-off task and fires with no task are
unchanged.
TestSpawnedTaskStartsWithGateOff becomes TestSpawnedTaskInheritsGateFromCaller.
New: TestSpawnTaskSeesCallerScopeThroughHarness and
TestTriggeredConversationInheritsGate.
The approval gate for asset targets now decides. It is opt-in per task and
per conversation (target_gate, default off). With the flag off nothing is
extracted, resolved, read or written, and every branch of the hook behaves as
before. No UI.
Extraction and normalisation (1.5)
- intercept.NormalizeTarget: lower case, trailing dot, userinfo, IPv6
brackets, IPv4-mapped IPv6, decimal / hex / octal / short IPv4, percent-
encoded host, punycode; port explicit, else the scheme default, else 0.
- intercept.ExtractTargets: WebFetch url exactly; Bash, shell_open, Monitor
(command) and shell_send (text) with the command-string rules (URL with a
scheme, IP, IP:port, host:port, CIDR; a bare name is not a target); every
string of the input for MCP tools, custom tools and any tool the gate does
not know. ExecuteExtraTool is unwrapped to the inner tool. The platform's
domain tools, the file tools and web_search have no target.
- The fields are decoded as the tools decode them (key case, last duplicate),
and a URL authority that reads differently quoted and unquoted yields both
readings.
Resolver, exemptions, gate decision (1.6)
- Resolver interface, system implementation (A and AAAA), 3 s per lookup,
positive answers cached 10 s per host with collapsed lookups, failures and
empty answers never cached, always fresh when a request is raised.
- Exemptions are exact host + port. Derived from the configuration (LLM
gateways, search backend in use, platform proxies): computed on every call,
never resolved, never voided. Stored (settings key
target_gate_exempt_hosts, seeded once with the package sources): resolved,
and void when any address is private, loopback, link-local, unspecified,
unique-local, multicast or a metadata address. An entry in both origins
counts as derived. A stored IP literal that can never be exempt does not
apply.
- GET / PUT /api/target-gate/exemptions and GET .../history. A change is
validated (ValidateExemptionHost on the normalised host) and appended to
target_gate_exemption_log in the transaction that stores it.
- Interceptor.CheckTargets: asset block rules (global and of the task) on the
name and its addresses, exemptions, resolution, pin lookup. A (host, 0) pin
covers every port; a (host, 443) pin covers 443 only and never a target
without a declared port; two pins are never merged. At most 5 unpinned
targets per call. Codes: TARGET_RULE_BLOCKED, TARGET_RESOLVE_FAILED,
TARGET_TOO_MANY, TARGET_GATE_UNAVAILABLE; none opens a request.
- With the upstream egress proxy set every pin is name-only at call time,
whatever name_only_effective says in its row.
- An approval never pins zero addresses: the pin writer refuses a request
that displays none (db.ErrTargetPinEmpty, 422 on the decide endpoint, the
row stays pending) and RequestTargets does not open one.
- Every block, and the first use per process of each exemption and of each
pin version, is a decided row (decided_by system, decision_source target).
The gate in the hook (1.7)
- guard.applyIntercept: a deny by a rule or by the judge blocks first, before
any target request exists; then CheckTargets; an allow is logged as before
and does not satisfy the gate; an ask and an unpinned target share one
request, and the classic ask is raised only when it was not merged. The
call is checked again after the decisions, at most three passes.
- The gate applies to every tool of a gated session, enabled for rules or
not, and to the target-only guard of the planner and the main agent, which
still records blocks only.
- Target blocks have their own model-facing text, built from the code and the
reason of the result.
Not in this change: rendered targets of custom tools (1.8) and redirect
targets (1.9). Subprocess paths stay advisory; the TestBypass_* tests document
what the gate does not see.
Storage, flag and request path of the approval gate for asset targets. The
gate itself is not wired into the tool hook yet (steps 1.5 to 1.7): with this
change nothing is extracted, resolved or blocked, whatever the flag says.
Schema and store (1.2)
- tasks.target_gate, conversations.target_gate (default false).
- intercept_pending.target_key and a partial unique index: one open request
per scope (task or conversation) and host:port.
- target_pins (no foreign keys: a pin outlives its task, its conversation and
its request row) and target_gate_exemption_log. All DDL is idempotent; no
CHECK constraints; the cascade on intercept_pending.conversation_id stays.
- db/target_pins.go: pin lookup with port precedence, list, revoke,
find-or-create of a request, and ResolveTargetRequest, which settles the
row, supersedes the old pin and writes the new version in one transaction.
- Deleting a task or an archive stub revokes its pins and settles its open
target requests in the delete transaction. DeleteConversation and
DeleteConversations revoke the pins of the removed conversations in the
same transaction as the delete.
Opt-in flag (1.3)
- Per task through the create and metadata-patch paths, per conversation
through create and patch; exposed in the task DTO. A retest conversation
takes the flag of the finding's task, never its pins. A spawned task starts
with the gate off.
- The task flag is atomic and read on every call through TargetScope.Enabled.
TargetScope now carries the Interceptor that owns the gate.
Request, wait, decide (1.4)
- pendingManager keeps a list of waiters per request; resolve wakes all.
- RequestTargets: find-or-create per target, one card per request, wait on
channel, poll, context and wait bound. A plain request is never settled by
a call that leaves. A request merged with an ask is bound to its creating
call and a timeout never allows it.
- Decide records the approver (subject of the session token) and the remote
address for every human decision, classic asks included. On a target
request "allowed" pins the displayed addresses; that is the only writer of
pins.
- A request raised while the global egress proxy is set says the approval is
name-only and its pin is stored with name_only_effective.
- Routes: GET/POST .../target-pins and .../target-pins/{pid}/revoke under
/api/tasks/{id} and /api/conversations/{id}; decision_source=target filter.
- ValidateExemptionHost: an exemption written as an IP literal may be private
unicast, never loopback, link-local or metadata, unspecified or multicast.
Step 1.1 of the S3b phase 1 plan. Wiring only: the approval gate for asset
targets is not enforced yet and no behaviour changes.
- Planner.Plan and MainAgent.Chat take a hooks parameter, like Worker.Execute.
Their sessions had no hooks, so nothing saw their tool calls before execution.
- guard.NewTargetOnly: the guard for those two agent kinds. It runs the
target-gate seam only: no intercept rule, no fallback judge, and no entry in
the audit ring (Guard.record is untouched and not called for these calls).
- Task.TargetGuard, built beside Task.Guard.
- intercept.WithTargetScope / TargetScopeFromContext: task id or conversation
id, an Enabled func and a wait bound, set at the five session entry points
(planner round, final planner round with wait bound 0, main-agent turn,
worker intent, conversation turn).
The seam allows every call and reads and writes nothing. No task or
conversation has an opt-in flag yet, so Enabled is nil (off) everywhere.
The IM renderers cut the summary of a single push at 120 characters and
listed at most 3 assets, as fixed constants. Both are now global settings
in the settings table, read by the delivery engine when it renders:
notify_summary_max_runes 0-4000, default 120
notify_assets_max_shown 0-50, default 3
0 means no truncation and an unset key keeps 120 / 3. PUT /api/settings
answers 400 for a value outside the bounds. The limits travel on
notify.Message, so the notify package still does not touch the database.
With no truncation, nothing bounded a single item in a Feishu card, and
Feishu rejects an oversized request as a whole. feishuCard now truncates
an item's text as the final backstop, the role TruncateBytes plays in
markdownBody for DingTalk and WeCom. feishuMaxCardBytes goes from 24000
to 16000: Feishu custom bots cap the request body at 20 KB, which the old
value exceeded.
Refs #168.
testLLM set cfg.UserAgent from the request (or the stored-profile
fallback) with no validation, while pgSaveProfile runs
agent.ValidateUserAgent and returns 400 on a bad value. A UA with a
control character could therefore reach the wire from the test button
though it is rejected on save. Call ValidateUserAgent on the final value
(after the stored fallback) and return 400 with the same message style,
mirroring pgSaveProfile; the fallback itself is unchanged.
ValidateUserAgent rejects every control character (r < 0x20 or r ==
0x7f), TAB included, so the header value stays a single printable token.
Extend the agent tests to cover NUL, US, DEL and TAB, and that a
printable slash-token value with spaces survives trimming unchanged. Add
a Postgres-free server test driving testLLM through an httptest endpoint:
a control-character UA is a 400 before any request leaves, a printable UA
reaches the wire verbatim.
Add an optional per-profile User-Agent that overrides the header on every
outbound LLM request. When empty (the default), the existing behaviour is
unchanged and the default User-Agent is left untouched.
The value is trimmed and validated: control characters / newlines (header
injection) and values longer than 256 bytes are rejected. Wired through the
profile config struct, the Postgres llm_profiles schema, the profile API/DTO,
the connection-test path, and the LLM profile form, following the existing
session-header field. Addresses part of #154; the fixed-full-URL mode is not
included.
- ARTEX_TRUSTED_PROXIES (comma-separated CIDRs, empty by default). When
the TCP peer is listed, the login limit reads X-Forwarded-For from the
right and keys on the first hop that is not a trusted proxy; otherwise
the header is ignored. Without it every client behind a reverse proxy
shares one bucket and an attacker can block the admin. Passed through
by docker-compose.yml and listed in .env.example.
- README: nginx example sends X-Forwarded-For; document the variable;
the binary's default listen address is 127.0.0.1:8787.
- reset-password.sh refuses passwords under 12 characters, counted in
characters like the server.
- POST /api/auth/login: 10 attempts per minute per TCP peer IP
(X-Forwarded-For ignored), then 429 with Retry-After. The block starts
at 1 minute and doubles on each attempt while blocked and on each new
block, capped at 15 minutes. A successful login clears the IP. No
per-account lockout. In-memory, idle entries swept after 15 minutes,
at most 10,000 IPs.
- New passwords (init, change-password, /setup page) need 12 characters.
Login does not check length, so existing 8-character passwords work.
- The binary's -addr default is 127.0.0.1:8787. The Docker image passes
-addr :8787 explicitly and is unchanged.
- GET /api/health returns {"ok":true} without a valid token; with one it
still returns service and version for the web UI top bar.
The server wrote jwt.key to the executable's directory, /app in the
image, so it lived in the container's writable layer and every container
rebuild generated a new key and signed all users out.
- ARTEX_JWT_KEY_DIR selects the key directory (default: config.BaseDir(),
unchanged for non-Docker installs). docker-compose.yml sets it to
/app/keys and bind-mounts ./keys there; the Dockerfile declares the
volume. The key stays out of data/, which the file manager browses.
- loadOrCreateJWTKey creates the directory (0700). When keyDir/jwt.key is
missing it migrates dataDir/jwt.key, then BaseDir()/jwt.key (mode 0600,
old file removed), otherwise generates a new key. An existing key is
never overwritten; a short one is now a startup error.
- .gitignore and .dockerignore exclude keys/.
- README and CHANGELOG updated; the 0.5.0 known issue is removed.
New accepted opts ...Option but never called them, so scrubSecrets stayed
false in the binary and ScrubSecretEnv was unreachable: the protection merged
in #136 never engaged. The unit test passed because it called the scrub
function directly instead of going through New.
Apply the options in the constructor and add a test that builds a Server
through New with WithSecretScrub and asserts the secrets are gone. Verified to
fail without the loop.
POST /api/llm/test and POST /api/llm/models take provider and base_url from
the form but fell back to the stored profile key (and, for the test, the
global key) when the form had none. Any base_url typed into the form, or
supplied by a crafted request, received the saved key.
The saved key is now used only when the request's provider and base URL
normalize to the ones it was saved with (sameLLMEndpoint, built on
agent.ConfigFrom). Otherwise the key must be entered in the form, and the
error says why.
POST /api/auth/init accepted any caller while auth.password_hash was empty,
so whoever reached port 8787 first could set the admin password and take the
instance. When no password is set at startup, the server now generates a
random one-time setup token and writes it to the server log. authInit
compares it in constant time, rejects a missing or wrong token with 403, and
clears it once the password is stored.
The setup page has a field for the token and api.initPassword sends it.
docker-compose publishes 8787 on 127.0.0.1 only, like 8788. README says where
to find the token.
The scrub ran in main after server.New returned, but New resumes every
non-terminal task before returning, and those loops start workers that run
shells. Move the scrub into New, right after both environment readers, and
start MCP discovery after it as well: both paths spawn child processes.
New takes an option rather than scrubbing unconditionally. The test suite
builds several servers in one process and still needs ARTEX_PG_DSN on later
calls, so only the binary opts in.
The update source is hard-coded to a repository that no longer exists and the release checksums are unsigned. Removes the selfupdate package and the /api/update/* endpoints, the top-bar check that queried GitHub on every page load, the Version & Updates card, and the exit-75 swap protocol in the start scripts.