A source now declares how its deliveries prove themselves: token (the
default, unchanged), hmac (HMAC-SHA256 over the raw body, per-source header
and prefix, secret sealed at rest and write-only) or none. The events module
verifies before the connector parses, answers 401 and logs the refusal.
The digest comparison is ported from internal/webhookguard/hmac.go and the
header-borne signature from internal/infrastructure/captureintake/receiver.go;
the old copies stay until their service goes. The seal package moves to
internal/platform/seal so events can seal the secret.
A conversation gets a persisted tainted flag (migration 00040), set the
first time output from a tool whose declaration trust level is not
"trusted" enters the turn, before that output reaches the model. The
flag is only ever set, never cleared, so later turns and resumed
processes inherit it. In a tainted conversation every non-read-only
action is raised to a floor of propose_and_wait, so the existing hold
path renders the approval card with the exact judged action; read-only
lookups still run, and a hard-rule refusal stays a refusal. An
unreadable taint state or a failed taint write fails closed. Trust
levels live on connectors.Declaration; unknown levels are untrusted.
Events (sources, triggers), workflows (list, document, YAML publish,
read-only canvas), runs with approvals and the decision trail, controls
(policies, guards, action classes), agent models, accepted extensions,
secret engines and credentials now read the binary's API. Capabilities
name every section, so pages the binary can't back yet (the visual
workflow editor, captures, mappings, bindings, the extension catalogue,
model aliases and the rest) stay in the tree, hidden. Adds
GET /api/credential-kinds for the credentials form.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Every /api operation is registered with Huma behind one admin-token
middleware, served at /api/openapi.json with a docs page at /api/docs,
and committed as api/openapi.yaml, which task check keeps current.
Resolve, target facts, invoke, lookup, deliver and parse record the protobuf wire messages as JSON on the context recorder the owning module supplies. Credentials are redacted and a body over 1 MiB keeps its prefix, full length and SHA-256. Judgement stores the resolve and target facts in the decision body and appends the executed invoke to it.
A new judgement module assesses one concrete action: hard rules refuse before
any guard runs; every guard sets a ceiling and the action runs at the lowest;
a guard that fails, panics or gives no ceiling holds it at propose and wait;
an advisory verdict is recorded and sets no ceiling. Reversibility and
Precedent (the fresh-install class default) are the first guards. Every
decision is recorded with its verdicts and class, and the run's trail lists
them. Agent tools now offer connector actions too: each call is assessed and
executed only as decided and only with the judged input, and a held action
comes back to the agent as a failed call naming its level and findings.
Tools gain Resolve, so an MCP server's declared effect reaches judgement.
Since the MCP migration made accepted.tools NOT NULL, the connector and secret
engine accept path never set it, so re-accepting any connector failed with a
not-null violation.
An investigation needs Prometheus's own metrics, and Prometheus publishes an
MCP server, so the connectors module now runs MCP servers as published instead
of wrapping them: a stdio binary in its own user and network namespace with an
egress allow-list, or an HTTP endpoint dialled through declared egress.
Acceptance names the transport, the tools it permits, and the credential slots
that resolve into the server's environment (stdio) or Authorization header
(HTTP). Each permitted tool is declared as a lookup, or as an action with its
effect; a tool the server lists but the acceptance doesn't name is never
offered, and a declared tool the server doesn't list fails the start rather
than going silently missing. Declared tools reach agents beside connector
lookups through the same declaration and the same call path, with the server's
own input schema, under a name no connector can take: an accepted name runs one
package.
A server that fails to start, exits or stops answering is reported by readiness
and reconnected with backoff. Tool calls carry a deadline, results are capped,
and an error result fails its call with the server's own text rather than the
run.
The investigate prompt now asks for a time-until-full figure from
`predict_linear` through the metrics server instead of an estimate from raw
points.
The old core MCP path stays for now: the legacy gateway and task containers
still read config.Tools.MCPServers, so it goes with them.
Credentials are typed (machine, age), write-only through the API and sealed
to the age identity in the bootstrap config. A field holds a value or names a
secret engine key; connectors and engines are accepted with credential slots
that resolve into their Configure secrets. The secret engine contract gains
secrets so an engine can receive its own key.
The agent model is an organisation setting in the new governance module.
Every model and tool call is a journal step, so a killed run resumes without
repeating either. Agents are offered connector lookups and read-only actions
only, until judgement exists.