87 Commits
Author SHA1 Message Date
sam f92c5b4536 chore(runtime): delete Kit and external agent harness support (#1877)
ci/woodpecker/pr/quality Pipeline was successful
Quality / check (pull_request) Canceled after 7m39s
Archie runs its own agent loop only. Removes Kit resolution and launch,
the harness runner and adapters, the setup terminal and sandbox view,
harness secrets and OAuth capture (table dropped in migration 0075), the
task-tools contract that only Kit containers reached, and the
sandbox-kit-spec dependency. Agent profiles carry image and tools only.
The egress proxy keeps the model-session path; the sandbox relay code
stays for #1867.
2026-11-12 21:29:05 +11:00
sam 0033933a58 perf(store): index the six unindexed foreign keys (#1860)
Postgres does not index the referencing side of a foreign key, so deleting or updating a parent row scanned the child table once per referencing row for the six keys that had no matching index: identity aliases and subjects, org memberships and agents, installed package requirements, and the task a handoff came from. Migration 0074 adds one supporting index per key; the foreign-key integrity checks on the restored snapshot pass, with every one of the 17 constraints now covered.
2026-11-11 02:49:37 +11:00
sam 4728c8c253 feat(usage): report model usage per org with identity and cost (#1855)
Model usage was recorded per call and never read, so an org could not see
what it spent. Add the read side:

- model_usage carries identity_id. Task calls take it from the task row,
  chat and curator calls from the calling principal carried in the usage
  record, so an outbox replay keeps its attribution. Existing task rows are
  backfilled from tasks.
- StatsService.UsageTotals sums the caller org's rows by UTC day, source,
  provider, model and identity. The org comes from the principal, with no
  org parameter, so cross-org reads do not exist.
- GET /api/usage prices tokens at the catalog's list rate, netting cache
  hits, and reports calls to unpriced models as tokens only. The published
  catalog view now carries each model's cost.
- A new org Usage page behind a read_usage grant: admin and owner only,
  with a 7, 30 or 90 day range, totals, the daily series, and model, source
  and identity breakdowns. The org shell keeps it on the caller's own org.
2026-11-11 02:15:29 +11:00
sam e3ffefdc7d fix(identity): use System as the one root identity and record who created every task (#1851)
deploy / plan (push) Successful in 25s
deploy / gate (push) Successful in 7m57s
deploy / agent-image (push) Successful in 21s
deploy / dist-zip (push) Skipped
deploy / archied-image (push) Successful in 2m12s
Drop the bot_user-derived root identity and the empty-identity sentinel. System
polls forge connections and owns instance-level work; a polled task stores the
polling identity, task_queued and work_request_submitted carry task id, actor
and principal, and an empty task identity is refused. Migration 0072 moves
empty tasks.identity, the identity-grants root key and the old root's
connection permits to System so a re-poll dedups.
2026-11-10 20:01:16 +11:00
sam e75d5c8ade fix(store): a restart re-queues a running task at its interrupted step and parks it after 3 restarts (#1841)
deploy / plan (push) Successful in 7s
deploy / archied-image (push) Canceled after 0s
deploy / agent-image (push) Canceled after 0s
deploy / dist-zip (push) Canceled after 0s
deploy / gate (push) Canceled after 4m4s
2026-11-09 14:36:08 +11:00
sam dc2c3b983f feat(workflow): a budget continuation resumes at the stopped step with its conversation (#1832)
deploy / plan (push) Successful in 7s
deploy / gate (push) Successful in 8m12s
deploy / agent-image (push) Successful in 28s
deploy / dist-zip (push) Skipped
deploy / archied-image (push) Successful in 2m11s
2026-11-09 12:44:01 +11:00
sam 3db75f9e28 fix(postgres): drop the removed schedules kind's apply status so archied is not degraded (#1834) 2026-11-09 09:41:50 +11:00
sam c30624a59e feat(workflow): a handoff starts a separate linked run of the next workflow (#1791)
Quality / check (push) Successful in 4m12s
deploy / gate (push) Successful in 4m42s
deploy / archied-image (push) Successful in 3m13s
deploy / agent-image (push) Successful in 11s
deploy / dist-zip (push) Skipped
2026-11-03 20:56:03 +11:00
sam 04ce0f9056 feat(identity): name, lifetime and last-used on personal API tokens 2026-10-27 15:37:44 +11:00
sam b8e080301c fix(postgres): let every org read a built-in mapping so a non-system org can arm a schedule trigger 2026-10-27 10:13:03 +11:00
sam b701f60a99 feat(scheduling): schedules and runs in the State Store with grid timing and run history
Schedules leave the schedules control-plane resource for State Store tables
behind a ScheduleService contract. A run is claimed per schedule slot, missed
slots fire once, the next run stays on the schedule's grid, cron is read in the
schedule-settings time zone, and failed runs are recorded and advance. Three
failures in a row raise a schedule.failing alert. A fired run records a
schedule-fired capture. Run history is pruned by age and count.

Removes the chat and spawn-task runners, the dangerous-payload guard and the
old schedules settings page.
2026-10-26 00:49:24 +11:00
sam 6ba6a595b8 feat(eda): capture schedule firings on a reserved source for bindings 2026-10-26 00:38:55 +11:00
sam 6a69fb760a fix(audit): harden export, scope history and derive feed targets 2026-10-25 22:39:22 +11:00
sam 274acfe6f5 fix(state): attribute identity events to the subject's single org 2026-10-25 21:35:19 +11:00
sam bc9b9952cd fix(state): classify audit coverage rows and keep unknown identity orgs private 2026-10-25 21:25:55 +11:00
sam 02a4ac0629 fix(state): harden the audit feed against audit findings 2026-10-25 21:23:52 +11:00
sam 390e89aefb fix(state): correct audit backfill scoping and policy write races 2026-10-25 20:59:04 +11:00
sam 8af19c79f4 feat(state): org-scoped audit trail and complete policy write audit 2026-10-25 20:49:20 +11:00
sam f2b8681611 refactor(workflow): stop assuming a human decides
An approval, ask or park can be resolved by an operator, a rule, a
classifier or automatically, so the vocabulary stops naming a person:

- step type human.approve -> workflow.approve
- park class needs_human -> needs_decision (ParkNeedsDecision)
- event kinds human_* -> operator_*
- alert kind task.needs_human -> task.needs_decision
- status grouping "needs you" -> "needs attention"

0061 rewrites the stored values, including the approval step type in the
workflow-definitions resource and each task's pinned definition, with the
digest recomputed so resumed tasks still match.
2026-10-25 17:31:16 +11:00
sam d699b9afba feat(workflow): approve a PRD as one structured document
The prd step returns the PRD as structured fields and human.approve takes
them as its document. The plan text every surface reads and the approval
card's layout are both rendered from that one document, so what the human
approves is what the implement workflow is handed. Replaces the separately
generated OpenUI view, which could drift from the plan.
2026-10-25 16:53:17 +11:00
sam 54a78175b4 feat(workflow): render a PRD's generated view on the plan approval card
agent.run gains views: the agent may return an OpenUI Lang view of its
answer through an optional view tool, read as steps.<id>.view. human.approve
stores it as the task's plan_view, and the approval card renders it with
the dashboard's own components, falling back to the plan text when there
is no view or it does not parse. The feasibility prd step uses it.
2026-10-25 16:33:04 +11:00
sam 75c1dbf83c fix(access): check the policy version before each decision
archied, archie-ui and the tool gateway re-read the stored policies on a 30s
ticker, so a forbid written seconds earlier did not stop a run. One
instance-wide revision now moves inside the transaction of every policy write,
and a decision reads it first and rebuilds the engine only when it moved.

A version read that fails, or a policy list that fails after a bump, refuses
with ErrChainUnavailable instead of standing on the last chain, so three call
sites now treat unavailable as availability rather than denial: dispatch leaves
the capture for the next cycle, a task start parks transient, and the tool
gateway reports the tool as unavailable.
2026-10-25 16:21:01 +11:00
sam 32b73c3a28 fix(store): post open chat task states once after the posted-status migration 2026-10-25 02:46:43 +11:00
sam a776de8b8f fix(archied): post each reportable status a chat task reaches, recorded by status not by flag 2026-10-25 02:42:31 +11:00
sam e8a60c4abd refactor!: waiting_human becomes waiting_approval, with the approval kind on each task 2026-10-24 23:45:20 +11:00
sam 8d0d707d3e fix(postgres): repin parked and queued tasks that still name the root identity 2026-10-19 01:45:48 +11:00
sam 9167c861d6 fix(postgres): run stored workflows that name the root identity as the seeded agent 2026-10-18 00:29:14 +11:00
sam 8aa59bdb51 fix(postgres): allow the agent org upgrade phase 2026-10-17 21:55:42 +11:00
sam a4da914994 fix(identity): emit RFC 4122 shaped stable IDs and migrate stored ones 2026-10-17 17:16:20 +11:00
sam 44a10943c3 feat(webui): link parked tasks to the page that fixes their blocker 2026-10-17 17:08:24 +11:00
sam 74699fcd21 refactor: remove personas
Personas were a misread requirement: SOUL is the chat identity and
identities are per action. The selected persona's prompt never reached the
model. Drops the resource, /personality, the chat selector, the settings
page and ChatService.SetPersona, and deletes stored persona documents.
2026-10-16 03:57:08 +11:00
sam 83aa17cf8b feat(identity): add browser setup and user accounts 2026-10-16 02:41:24 +11:00
sam 33ccfb2505 fix(daemon): run-time workflow check sees the binding event's signature and address
Issue: archie-core-bmro8
2026-10-13 15:10:31 +11:00
sam 8a3ba32c5a Feat/operator alerts (#1635)
* feat(state): durable operator alerts owned by the State Store

Alerts dedup on org+kind+subject, resolve themselves when a condition
clears, and are raised, resolved, dismissed and listed through the state
contract. Task transitions raise needs-human and failed alerts in the same
transaction; the store checks service health and prunes closed alerts.
Each raise and resolve queues a webhook delivery whose attempts are kept.

* feat(notify): deliver operator alerts to the webhook; raise config and release alerts

notify.webhook now receives raised and resolved alerts above a severity
threshold with HMAC-signed payloads, bounded retry and every attempt recorded
on the alert. archied raises model alias, provider and credential-binding
alerts from the effective configuration, an update alert from the release
channel, and a credential-rejected alert from the egress resolver. The
dashboard API lists and dismisses the caller's org's alerts.

* feat(ui): alert bell in the top bar and webhook alert settings

The bell lists the org's open alerts by severity with dismiss and a link per
alert, and refreshes from the event stream. The task settings page gains the
webhook severity threshold and signing secret. ui/dist rebuilt.
2026-10-12 03:30:46 +11:00
sam b8f8f9a272 fix(postgres): allocate synthetic issue numbers from a per-repo counter
InsertChatTask and EnqueueCallTask computed MAX(issue_number)+1, so
concurrent inserts in one repository took the same number and the second
failed on the unique key. Both now bump a chat_issue_numbers row, seeded
from the existing maxima, whose row lock serializes allocation.
2026-10-11 14:21:27 +00:00
sam f5a6fa5b32 fix(postgres): number the outcome-posted migration after package updates 2026-10-11 04:04:45 +11:00
sam cac0f43b58 Merge branch 'work/mqti13'
# Conflicts:
#	internal/contracts/state/v1/state.pb.go
2026-10-11 04:03:52 +11:00
sam 805364f201 feat(archied): post a chat task's outcome back to its chat 2026-10-11 04:03:30 +11:00
sam b227df3c3f feat(storepkg): widening updates wait for approval, rollback restores the previous pin 2026-10-11 03:48:31 +11:00
sam e988d3cc40 fix: close cross-org writes and move existing rows into their owner's org 2026-10-10 20:31:54 +11:00
sam 0fcf792db3 feat(gateway): chat sessions belong to one org
Sessions carry an org; listing, reading and writing a session, its
messages and its turns are limited to the caller's org. Another org's
session reads as not found and cannot be overwritten or appended to.
Sessions from before orgs belong to the system org.
2026-10-10 18:57:26 +11:00
sam e4ea2eaabc feat(statestore): record a usage row for every model call
Task agent runs and MCP sampling record against their task (billed to the
task's org); gateway chat turns, titles, sampling and curators record
against the system org. Rows carry source, workflow, step, alias,
provider/model and tokens for later chargeback.
2026-10-10 18:32:43 +11:00
sam 7680dc5079 feat(presence): expose service dependency checks 2026-10-09 09:57:56 +11:00
sam 3c99f69fee fix(workflow): key a call on its callee, and free a terminal child's key
A call's key was the step's declared path alone, so one task running a
second definition at a colliding path -- workflow.handoff, human.approve
and a definition re-pin all change it -- was handed the first callee's
child, waited on the wrong workflow and published its outputs. The key
now carries the callee too, while the step row and the call's events stay
named and staged for the call site, so a rail pairing an event with its
step still attaches them.

A child that reached dead or closed_wont_do can never do the work and
holds no result the caller can use, so it leaves the call key's index and
frees the key for the caller's retry to start a fresh child. parked and
rejected keep theirs: a parked child is recoverable by retrying it, and a
rejected child ran.

The id a call step must declare is now refused where the definition is
compiled, which is also the path a pinned definition takes, so an id that
spells a step type is accepted again and no stage needs a heuristic.
2026-10-08 22:31:07 +11:00
sam b7fd13923a fix(workflow): key a workflow.call on its declared step path
A call site had no durable identity, so any retry of it -- a lost
EnqueueCallTask reply, a failed call-step record, a requeued attempt that
re-runs the stage -- started a second child and its work happened twice.

The call step's declared path within the run is that identity: wrapStep
sets it (branch-qualified, restored on return), a workflow.call step must
declare an id so two call sites cannot share one, and the key crosses
EnqueueCallTaskRequest.call_key into tasks.call_key, which EnqueueCallTask
upserts on so the retry returns the child already started.
2026-10-08 22:31:07 +11:00
sam 8d73280633 feat(telegram): /stop stops the chat agent, else the tasks it created, or one task; delete the dead approval actions 2026-10-06 01:12:28 +11:00
sam 0fc8535d87 fix(eda): name a source's delivery header in the dashboard instead of a built-in sender list 2026-10-06 00:54:20 +11:00
sam a355f13a03 fix(eda): dispatch each webhook delivery once, however often the sender retries it 2026-10-06 00:51:45 +11:00
sam 0f0795e6f1 feat(eda): name and delete capture sources; an armed binding blocks the delete 2026-10-05 21:05:58 +11:00
sam 0acf7981d9 feat(access): the dashboard operator is a person owning org-sys 2026-10-05 20:33:22 +11:00