Agent Workflows

Use this guide to review the thirteen cross-agent workflows that the pantheon composes into product-level capabilities. Each workflow names its participating agents, its trigger, its end-to-end sequence, and its exit criteria. Every workflow ships in observation mode first (agent-pantheon-implementation.md § Wave 7) and is promoted per-workflow after Wave 8 measures its KPIs.

Scope: the workflows are customer-agnostic. Concrete resource names in examples are placeholders (generic-scope.instructions.md).

Contract: every step is a pub/sub event on a schema-checked topic (see agent-pantheon.md § 6.1). No workflow uses direct RPC between agents. human approval steps go through Var; audit goes through Saga. There are no shortcuts.

Machine-readable form. Shipped executable workflows live under rule-catalog/workflows/ or a package-owned equivalent. The exact-id boundary inventory currently binds predictive-scale and the optional cost-aware-remediation package workflow. The other eleven design workflows remain metadata-only agent choreography and do not imply one file per section. Similar but differently named Process workflows are independent catalog entries. The schema, Process ObjectType, and compile-to-Runbook wiring are defined in process-automation.md.

Implementation status

Implementation scope

Area	State	Evidence	Notes
Thirteen-workflow metadata registry	implemented	services/core-control-plane/src/fdai/agents/_framework/workflows.py; services/core-control-plane/tests/agents/test_wave7_workflows.py	All registered workflows default to shadow; the registry is metadata and does not by itself prove a deployed end-to-end workflow.
Executable observation mode trace references	implemented	services/core-control-plane/tests/agents/test_wave7_workflows.py; services/core-control-plane/tests/composition/test_readiness_service.py; services/core-control-plane/tests/core/test_control_loop_operator_request.py; services/core-control-plane/tests/agents/test_detection_readiness.py	Focused tests cover the registered trace paths. They are implementation evidence, not retained operational traces.
Published workflow sequence diagrams	validated	docs/diagrams/fdai-agent-workflows-*.diagram.yaml; tools/architecture-diagrams/test/agent-workflows.test.ts; exact-SHA CI and Pages runs; live bilingual geometry checks	All twelve published diagrams show complete sender and receiver names plus the typed message in centered bilingual cards. This presentation adds no direct call, workflow state, authority, or promotion evidence.
Metadata-to-executable catalog boundary	implemented	config/workflow-promotion-verdicts.json; scripts/quality/architecture/check-workflow-promotion-verdicts.py; focused checker tests	Exact workflow ids bind two metadata workflows to executable catalog entries; eleven remain metadata-only. All 14 other executable YAML workflows are explicitly classified as process-catalog-only. Canonical YAML digests bind triggers, steps, default modes, promotion gates, and all other declaration fields. Similar names do not create a mapping, and the inventory grants no mode or authority.
Measured promotion gates	not-started	Promotion thresholds in this document and services/core-control-plane/src/fdai/agents/_framework/workflows.py	No retained evidence demonstrates the required observation mode durations, KPI baselines, or per-workflow gate results.
Per-workflow promotion decision inventory	implemented	config/workflow-promotion-verdicts.json; scripts/quality/architecture/check-workflow-promotion-verdicts.py; focused checker tests	Each of the 13 metadata workflows has an exact-definition decision: 12 are deferred for named missing operational evidence, and retrospective what-if is permanently observation mode. The inventory records no operational evidence and changes no mode or authority.
Enforce-mode promotion	not-started	default_mode="shadow" in services/core-control-plane/src/fdai/agents/_framework/workflows.py	Promotion remains independent per workflow; retrospective what-if is inherently observation mode and is not eligible for enforcement.

Implementation history

Date	State	Change	Evidence	Remaining
2026-10-15	implemented	Recorded and gated the non-1:1 metadata-to-executable boundary for all 13 workflows. Exact ids bind two entries; the other eleven are metadata-only, and similar-name substitution or duplicate exact ids fail closed.	current change; workflow decision config/schema/checker; 10 focused checker tests.	Retain operational observation mode duration, KPI baseline, policy-escape, and trace evidence before any eligible promotion.
2026-10-15	implemented	Extended the boundary across all 16 executable YAML workflows. Fourteen independent process-catalog entries are explicit, and canonical definition digests invalidate the inventory when triggers, steps, default modes, promotion gates, or other declaration fields drift.	current change; workflow decision config/schema/checker; 15 focused checker tests.	Retain operational observation mode duration, KPI baseline, policy-escape, and trace evidence before any eligible promotion.
2026-10-01	implemented	Recorded the implemented workflow steps from the I12 batch in the workflow registry, trace assertions, planned-agent exit conditions, and doc parity tests without changing any workflow mode.	current change; services/core-control-plane/src/fdai/agents/_framework/workflows.py; services/core-control-plane/tests/agents/test_wave7_workflows.py; docs/roadmap/agents/agent-workflows.md; docs/roadmap/agents/agent-workflows-ko.md; focused Wave 7, Pantheon doc parity, localization, roadmap, link, punctuation, readable-Hangul, and core-import checks.	Retain operational observation mode duration, KPI baseline, policy-escape, and promotion evidence before any enforce-mode change.
2026-09-16	implemented	Recorded one fail-closed decision for every metadata workflow without treating implementation traces as operational measurements. Exact catalog and definition digests make a new or changed workflow fail the inventory gate until its decision is reviewed again.	current change; config/workflow-promotion-verdicts{,.schema}.json; scripts/quality/architecture/check-workflow-promotion-verdicts.py; focused checker tests passed 7 cases.	Retain the named runtime duration, KPI baseline, guard-regression, and policy-escape evidence before replacing any deferral. No workflow mode changed.
2026-08-20	validated	Retained exact-source CI, Pages deployment, and live bilingual geometry evidence for the corrected workflow diagrams. Every one of the 24 deployed SVGs exposes a message body for every node, centers the sequence with zero delta, and has zero text overflow or node overlap; the English and Korean routes also have no page or diagram-host overflow at desktop, constrained-desktop, or mobile widths.	Commit c22ea624b; CI run 32336843459; Pages run 32336843527; live 1440x900, 993x641, and 390x844 checks.	None for the published sequence-diagram regression. Runtime promotion evidence remains separately open.
2026-08-20	implemented	Corrected the published sequence presentation after live review found that every workflow collapsed into a narrow left-aligned actor chain, hid typed messages from the visible cards, and truncated return-arrow senders such as Njord. Sequence cards now expose bounded message bodies, center the ordered chain, and preserve complete participant aliases.	current change; twelve bilingual workflow specs and mirrored assets; 95 diagram compiler tests, typecheck, artifact freshness, 35-pair public migration, 10 focused site contracts, and direct EN/KO geometry checks passed with zero text overflow or node overlap.	Retain exact-source Pages deployment evidence before closing the visual regression. Runtime promotion evidence remains separately open.
2026-08-13	implemented	Adopted the implementation ledger and reconciled the workflow inventory with the metadata registry and focused observation mode tests. Earlier implementation provenance was not reconstructed.	current change; focused workflow tests	Complete catalog projection where required, retain operational observation mode evidence, and evaluate promotion gates independently.

Remaining work

• Record and gate the exact-id catalog boundary for all 13 design workflows and all 16 executable YAML workflows: two exact ids are joined, eleven design workflows remain metadata-only, and fourteen executable workflows are explicitly process-catalog-only.
• Retain per-workflow observation mode-duration, KPI-baseline, policy-escape, and trace evidence from an operating environment.
• Record one exact-definition promotion decision for every metadata workflow. The current inventory defers 12 workflows for named missing evidence and keeps retrospective what-if permanently observation mode; it grants no authority.

0. Workflow shape

Every workflow declaration follows the same structure:

• Purpose - what business capability the workflow delivers.
• Trigger - the event or schedule that starts the flow.
• Agents - primary and supporting, with role labels.
• Sequence - a static SVG diagram showing typed-port messages.
• Exit criteria - measurable conditions for observation mode trace success.
• Promotion gate - the KPI thresholds required for enforcement mode.
• Anti-scope - what the workflow deliberately does not do.

Workflows do not add new ontology types or ActionTypes; they consume the existing catalog under rule-catalog/action-types/ and the object types under rule-catalog/vocabulary/object-types/. A workflow that needs new types is a signal to open an upstream doc PR first.

1. Cost-aware fix

Purpose. Every SRE fix carries an attached cost impact so the decision reflects both reliability and finance. Prevents automation from saving one dollar of on-call time by spending ten dollars of compute.

Trigger. Heimdall publishes object.drift (declared vs actual state mismatch) or object.anomaly on a resource with an existing rule match.

Agents. Heimdall (initiator), Njord (cost advisor), Forseti (judge), Thor (executor), Saga (auditor).

Current executable trace agents. Njord, Forseti, Thor, Saga. Planned workflow agents. Heimdall.

Exit criteria.

• Current trace assertion cost_advisory_measured: Njord returns a measured signed estimate for the proposed action.
• Current trace assertion cost_ceiling_blocks_auto: the focused trace verifies the ceiling rule that sends over-ceiling cost impact to human approval.
• Current trace assertion cost_annotation_attached_to_verdict: Forseti attaches bounded Njord cost evidence to the object.verdict payload.
• Current trace assertion terminal_action_run_audited: the accepted auto decision reaches Thor and Saga records the terminal object.action-run.
• Planned exit condition: Heimdall must feed object.drift or object.anomaly as the workflow initiator into the current Njord -> Forseti -> Thor -> Saga trace before this workflow can claim initiator coverage.

Promotion gate. 14 days observation mode; Njord cost forecast MAPE < 20% on this workflow’s audit sample; zero missing cost_annotation on fixes.

Anti-scope. Not a budget enforcement (Njord already emits CostAnomaly for that separately); this only annotates SRE actions with cost.

2. Predictive scale

Purpose. Scale proactively before Freyr’s forecast trips a threshold instead of reactively after Heimdall detects saturation.

Trigger. Freyr recurring forecast run (hourly). When the forecast predicts threshold breach within fork_config.predictive_horizon (default 2 hours).

Agents. Freyr (initiator), Heimdall (early-signal cross-check), Njord (cost check), Odin (arbitration if cost blocks scale), Forseti, Thor.

Current executable trace agents. Freyr, Heimdall, Njord, Odin, Forseti. Planned workflow agents. Thor.

Exit criteria.

• Current trace assertion forecast_leads_reactive_baseline: the capacity forecast crosses the scale threshold more than 30 minutes before the paired reactive baseline.
• Current trace assertion false_positive_baseline_checked: a paired below-threshold series does not recommend scale-up.
• Current trace assertion arbitration_request_on_cost_conflict: a Njord cost-block signal on the same resource creates exactly one Forseti arbitration request for the capacity/cost conflict.
• Current trace assertion recurring_sample_maps_to_shadow_scale_verdict: Freyr recurring sampling publishes a capacity forecast that Forseti maps to an ops.scale-out or ops.scale-in observation mode/decision requiring human approval.
• Planned exit condition: Thor must dispatch a governed scale ActionType and retain independent effect observation before the workflow can claim executable scale coverage.

Promotion gate. 30 days observation mode; Freyr forecast MAPE < 15% on this workflow’s samples; false-positive scale rate < 5%.

Anti-scope. Not autoscale rules (existing platform autoscale keeps running); this triggers deliberate scale actions attributable to Freyr’s forecast.

3. DR drill orchestration

Purpose. Regular disaster-recovery rehearsal without waiting for a real incident. Verifies Vidar’s rollback paths, DR failover mechanics, and observability all still work.

Trigger. Loki schedule (weekly by default, fork-configurable).

Agents. Loki (planner), Forseti (judge), Var (approver), Vidar (execution), Heimdall (observation), Norns (learning), Saga.

Current executable trace agents. Loki, Thor, Vidar, Saga. Planned workflow agents. Forseti, Var, Heimdall, Norns.

Exit criteria.

• Current trace assertion blast_radius_capped: Loki caps a requested experiment to the configured impact scope and marks human approval as required.
• Current trace assertion recurring_scheduler_publishes_hil_drill_window: Loki’s deterministic scheduler emits one complete always-human approval drill proposal for a due window.
• Current trace assertion proposal_audited: Saga audits the Loki object.chaos-experiment proposal.
• Current trace assertion vidar_dr_contract_gates_failover_recovery_time: Vidar accepts a typed DR failover contract, Thor waits for that contract before executor I/O, and Vidar publishes a DR outcome with recovery time.
• Planned exit condition: Forseti and Var must route the drill through a governed judgment and human approval, Heimdall must publish the independent observation, and Norns must compare the previous drill baseline and raise one candidate for any MTTR degradation above 20%.

Promotion gate. 3 successful drills in observation mode; drill duration < declared budget; zero unplanned production side-effects (measured by Heimdall’s impact scope audit).

Anti-scope. Not real DR - this is rehearsal only. Real DR failover uses the same Vidar action type but with a different trigger (incident-classified emergency).

4. Override -> Discovery

Purpose. Every human override of a rule decision becomes a signal for rule refinement. Frequent overrides on the same rule mean the rule is either wrong, over-scoped, or missing a critical exception.

Trigger. Var records Approval where the operator’s decision differs from Forseti’s proposed decision (approve on deny, reject on auto, etc.).

Agents. Var (initiator), Saga (aggregator), Norns (learner), Mimir (rule accountable owner).

Current executable trace agents. Var, Saga, Norns, Mimir. Planned workflow agents. None.

Exit criteria.

• Current trace assertion approval_override_source: Norns learns from Var-owned object.approval rejections that carry an override signal.
• Current trace assertion deduped_rule_candidate: recurring decisions for the same action produce exactly one inert RuleCandidate.

Promotion gate. 60 days observation mode; override-to-candidate conversion rate matches expected pattern (i.e., not every override becomes a candidate); false-candidate rate < 10% (Mimir reject rate).

Anti-scope. Does not auto-modify rules. Every candidate goes through Mimir’s normal promotion pipeline.

5. Security escalation

Purpose. Formalizes the privilege-escalation monitoring flow from agent-pantheon.md § 9 as a first-class workflow with promotion gate.

Trigger. Forseti emits object.security-event with type: privilege_escalation_attempt.

Agents. Forseti (initiator), Heimdall (correlator), Odin (critical severity path), Var (admin notification delivery via ChatOps), Saga.

Current executable trace agents. Forseti, Heimdall, Var. Planned workflow agents. Odin, Saga.

Exit criteria.

• Every RBAC-deny produces exactly one SecurityEvent.
• Severity classification is deterministic (counter + table only).
• Current trace assertion duplicate_same_user_action_upserts_one_card: same-user same-action alerts collapse to one card with an incrementing counter.
• Current trace assertion per_user_rate_limit_blocks_sixth_card: the sixth high-severity card for the same user in the rolling hour is held.
• Current trace assertion critical_pattern_pages_admin: the critical multi-action pattern emits an admin card.
• Planned exit condition: Odin must publish the critical escalation path and Saga must retain replayable audit evidence before the zero-false-negative gate can be evaluated.

Promotion gate. 30 days observation mode; zero false negatives on injected critical patterns; false-positive rate on high < 5%.

Anti-scope. Does not implement permission-upgrade flow (that is future work, see pantheon § 9.5).

6. Handoff -> Capability

Purpose. Every unhandled request (Handoff) is a capability gap. Repeated handoffs of the same fingerprint should convert into new rules or new agent capabilities.

Trigger. Saga writes object.issue (via escalate_to_github_issue action). Norns aggregates by fingerprint.

Agents. Saga (initiator), Norns (aggregator), Mimir (rule accountable owner), Bragi (updated on capability delivery).

Current executable trace agents. Saga, Norns, Mimir. Planned workflow agents. Bragi.

Exit criteria.

• Handoff fingerprint occurrence count monotonically tracked.
• Current trace assertion candidate_deduped: RuleCandidate emits when the threshold is exceeded (dedup: one candidate per fingerprint per rolling window).
• Current trace assertion failed_promotion_keeps_issue_open: a refused Mimir promotion leaves the Saga issue open.
• Current trace assertion norns_quiet_window_signal_is_inert: Norns emits an inert issue-close eligibility signal only after the quiet window and never mutates the Saga issue.
• Current trace assertion promotion_evidence_closes_after_clean_window: Mimir promotion evidence plus a 24-hour clean regression interval lets Saga close the issue with the promoting PR reference.
• Planned exit condition: Bragi must deliver the closed capability in the operator briefing so the workflow is visible after Saga closes the issue.

Promotion gate. 90 days observation mode; conversion rate (handoff -> promoted rule) baseline captured; false-close rate < 2%.

Anti-scope. Does not auto-write rule text. Candidates carry evidence and a proposed shape; Mimir + humans review and refine.

7. Agent health degradation

Purpose. When an agent itself is failing, the system detects it, adjusts portfolio priority, and briefs operators - not silently degrading and only surfacing when a workflow breaks.

Trigger. Heimdall recurring agent-health probe (per-minute heartbeat + KPI compare vs baseline). Detects heartbeat gap, high error rate, or KPI drift.

Agents. Heimdall (detector), Odin (portfolio re-planner), Bragi (operator briefing), Saga.

Current executable trace agents. Heimdall, Odin. Planned workflow agents. Bragi, Saga.

Exit criteria.

• Current trace assertion odin_arbitrates_degradation_priority: Odin can arbitrate a health-degradation priority conflict deterministically.
• Planned exit condition: Bragi must deliver the operator briefing within 60 seconds and Saga must retain the degradation audit while Heimdall probes every agent at the declared frequency and the degradation policy matches pantheon anti-patterns table.

Promotion gate. 30 days observation mode; every declared degradation policy tested by injected failure at least once; briefing latency p99 < 60s.

Anti-scope. Not self-heal - Heimdall does not restart failing agents. Recovery is a separate operator action (ideally through Vidar if a rollback path exists).

8. Judgment coherence audit

Purpose. Verifies that Forseti’s decisions remain consistent over time - the same input should produce the same decision, absent rule change. Catches model drift, rule catalog corruption, and non-determinism bugs.

Trigger. Forseti recurring self-test (daily). Samples recent decisions, re-runs them, compares.

Agents. Forseti (self-tester), Muninn (audit sample), Norns (drift analyzer), Mimir (reviews if drift is caused by rule change), Saga.

Current executable trace agents. Forseti, Saga, Norns. Planned workflow agents. Mimir.

Exit criteria.

• Current trace assertion audit_sample_replayed: Saga provides the replayed audit sample used for the comparison.
• Current trace assertion forced_mismatch_creates_one_candidate: a forced mismatch is classified into exactly one candidate and one alert record by the focused trace harness.
• Planned exit condition: the recurring daily job, Muninn sample fetch, and Mimir rule-change review must exist before the 15-minute budget and false-drift-alert promotion metrics are evaluated.

Promotion gate. 60 days observation mode; mismatch rate baseline captured; false-drift-alert rate < 5%.

Anti-scope. Does not roll back rule changes automatically. Any alert is investigatory.

9. Rollback rehearsal

Purpose. Proactively test that rollback paths declared in ActionType rollback_contract actually work. Prevents detected issue out at incident time that rollback is broken.

Trigger. Loki schedule (monthly). Picks a subset of ActionTypes based on fork_config.rollback_rehearsal_scope.

Agents. Loki (planner), Forseti (judge), Var (approver), Vidar (rehearser), Heimdall (observer), Saga.

Current executable trace agents. Loki, Vidar, Saga. Planned workflow agents. Forseti, Var, Heimdall.

Exit criteria.

• Current trace assertion blast_radius_full_blocks_overlap: Loki refuses a second proposal when the rehearsal impact scope reservation is full.
• Current trace assertion proposal_audited: Saga audits the accepted Loki rehearsal proposal.
• Current trace assertion vidar_records_non_mutating_rehearsal_receipt: Vidar records a bounded non-mutating rehearsal receipt for the bound rollback contract.
• Planned exit condition: Forseti and Var must route the rehearsal through governed judgment and human approval, and Heimdall must compare post-rollback state with the pre-mutation baseline and raise a candidate for any deviation.

Promotion gate. 3 successful rehearsals per ActionType before that type is eligible for enforcement mode outside observation mode. Rehearsal cadence enforced by Loki schedule.

Anti-scope. Not production rollback (that uses the real path when Vidar responds to a real failure).

10. Retrospective what-if

Purpose. Given a past incident (in audit log), re-play judgment under different rule configurations to answer “if we had had this rule at the time, would the incident have been prevented?” - crucial for Mimir’s rule promotion decisions.

Trigger. Manual (operator via Bragi) or scheduled (post-incident).

Agents. Saga (data source), Forseti (re-judge), Norns (delta analysis), Mimir (rule evaluation), Bragi (report).

Current executable trace agents. Saga, Forseti. Planned workflow agents. Bragi, Norns, Mimir.

Exit criteria.

• Replay is judge-only (never re-executes).
• Current trace assertion overlay_rejudgment_reproducible: replaying the same audit slice under the same scoped judgment overlay produces the same what-if decision.
• Current trace assertion no_action_run_published: the replay publishes no object.action-run.
• Current trace assertion versioned_what_if_disagreement_evidence_inert: Forseti publishes versioned retrospective what-if evidence with bounded disagreement reasons and a shadow_only ceiling.
• Planned exit condition: Bragi report generation, Norns delta analysis, and Mimir rule evaluation must consume the replay result before this is more than a judge-only what-if trace.

Promotion gate. Not applicable (this workflow is inherently observation mode - it never executes changes).

Anti-scope. Does not modify Saga audit log. Overlay is a read-time projection.

11. Operational readiness handoff

Purpose. Gate the dev-to-ops boundary: before a dev-owned scope becomes the operations team’s responsibility, review its accumulated governance, security, RBAC, and reliability posture and return one decision (clear / needs_review / blocked). Catches gaps a per-change review misses - an over-privileged workload identity, a guest holding Owner, missing backup - that no single diff introduced. Full design: operational-readiness.md.

Trigger. Huginn normalizes an ownership_transfer signal (a handoff PR label, a lifecycle-stage: handoff tag, or an operator request_ops_handoff) carrying the target scope, submitter, and target environment.

Agents. Huginn (collector), Mimir (applicable rule set), Forseti (judge / ReadinessReport), Var (human approval approver on blocked handoff + proposed fixes), Thor (executor of approved fixes), Saga (auditor).

Current executable trace agents. Forseti, Var, Thor, Saga. Planned workflow agents. Huginn, Mimir.

Exit criteria.

• Current trace assertion composition_handoff_blocks_on_critical_finding: the composition readiness service blocks handoff on a critical detected issue and routes approved fixes through the governed path.
• Planned exit condition: Huginn-owned ownership_transfer ingress and Mimir-owned applicable rule selection must feed the Pantheon trace before the workflow can claim one ReadinessReport per transfer signal.

Promotion gate. 30 days observation mode per environment; zero false negatives on injected critical identity patterns; false-positive rate on blocking detected issues < 5%.

Anti-scope. Does not execute fixes itself (proposes only; RBAC fixes route to human approval via remediate.right-size-role). Does not define the environment model (consumes scope-expansion.md). Not a per-deploy check (that is deployment-preflight.md).

12. Scheduled governed Python task

Purpose. Run an immutable generated Python artifact on one inventory-selected GPU VM without giving the authoring surface a VM identity or accepting shell text.

Trigger. Strict five-field cron schedule materialized by the scheduler with a target Resource and PythonTask artifact binding.

Agents. Bragi owns authoring translation, Forseti owns the risk decision, Var owns Owner human approval approval, Thor owns Managed Run Command execution, and Saga owns the audit record. The current runtime maps these responsibilities to the authoring API, scheduler plus EventIngest, unified safety check, human approval resume coordinator, and tool executor. The optional Pantheon consumer remains a observation mode observer and does not execute the proposal.

Current executable trace agents. Forseti, Var, Thor, Saga. Planned workflow agents. Bragi.

Exit criteria. Current trace assertion control_loop_owner_approval_reaches_runner: the raw control-loop proposal reaches the VM runner after Owner approval. Planned exit condition: Bragi must route the authoring translation into the scheduled proposal, and the scheduler-owned cron materialization, artifact recheck on every guest invocation, active compute.vm target binding, GPU capability check, retry idempotency, remote-cancel path, and terminal audit must all be present before the scheduled workflow gate can be evaluated.

Promotion gate. 14 days and 30 observation mode plans; accuracy >= 99%; zero policy escapes; explicit Owner review before FDAI_VM_TASK_ENFORCE=1.

Anti-scope. Does not provision VMs, install packages or drivers, accept shell commands, pass source through the event bus, or bypass the safety check.

13. Detection readiness assurance

Purpose. Reduce per-target detection-pipeline signals across six dimensions (discovery, collector configuration, telemetry, detector binding, pipeline coverage, and action governance) into one authoritative readiness decision, so that incomplete, malformed, or stale detection coverage never lets an auto-execution decision rise above shadow for that target.

Trigger. detection.readiness.observed events arriving on the raw ingress topic, one per readiness dimension per target and pass.

Agents. Huginn ingests and deduplicates raw observations by idempotency key. Heimdall reduces a completed pass’s six dimension observations for a resource into a decision (ready, partial, blocked, stale, unauthorized, or unknown) and publishes object.drift; a still-in-progress pass is never replaced by an overlapping or later pass until it completes, and a completed drift is never re-emitted for a new pass at the same resource. Muninn persists exactly one durable snapshot per resource, rejecting duplicate or out-of-order deliveries by generated_at, and Saga audits the resulting state-snapshot transition. Forseti records the decision on its own drift stream without creating a decision, then demotes any subsequent auto-execution decision for that resource to hil while its readiness ceiling remains below the required level. Bragi is planned for operator presentation; the current runtime does not yet route detection-readiness traffic through it.

Current executable trace agents. Huginn, Heimdall, Muninn, Forseti, Saga. Planned workflow agents. Bragi.

Exit criteria. Current trace assertion readiness_reduced_in_shadow: a complete raw pass reduces to one observation mode readiness drift. The focused detection readiness suite also verifies malformed observations, Huginn replay deduplication, overlapping partial passes, Muninn stale-snapshot rejection, Saga audit, and Forseti demotion while the recorded readiness decision remains below the required ceiling. Planned exit condition: Bragi must route detection-readiness presentation into the operator-facing briefing before this workflow can claim operator presentation coverage.

Promotion gate. 30 days observation mode per target; zero false-ready snapshots; stale-detection p99 < 15 minutes.

Anti-scope. Does not execute or verify any action itself, does not create a risk decision directly (only demotes decisions raised by the normal event path), does not treat a partial or malformed observation set as ready, and does not bypass Muninn’s ordering/dedup checks for late or duplicate snapshots.

14. Workflow catalog summary

#	Name	Trigger	Primary agent	Enforce prerequisite
1	Cost-aware fix	Drift / anomaly	Heimdall + Njord	Cost forecast MAPE < 20%
2	Predictive scale	Freyr forecast (hourly)	Freyr	Forecast MAPE < 15%, FP < 5%
3	DR drill orchestration	Loki schedule (weekly)	Loki	3 observation mode drills clean
4	Override -> Discovery	Var override event	Var	Conversion rate baseline
5	Security escalation	Forseti RBAC deny	Forseti	Zero critical FN, FP < 5%
6	Handoff -> Capability	Saga issue creation	Saga	Conversion baseline, FC < 2%
7	Agent health degradation	Heimdall probe	Heimdall	Every degradation tested
8	Judgment coherence audit	Forseti self-test	Forseti	Drift-alert FP < 5%
9	Rollback rehearsal	Loki schedule (monthly)	Loki	3 rehearsals per ActionType
10	Retrospective what-if	Operator or post-incident	Bragi	(inherently observation mode)
11	Operational readiness handoff	ownership_transfer signal	Forseti	30d observation mode/env, zero critical FN, FP < 5%
12	Scheduled governed Python task	Strict cron schedule	Forseti + Thor	30 plans, >= 99% accuracy, zero escapes, Owner human approval
13	Detection readiness assurance	detection.readiness.observed	Heimdall	30d observation mode/target, zero false-ready, stale p99 < 15m

Next steps

To learn about	Read
The pantheon roles referenced above	agent-pantheon.md
The wave plan that lands each workflow	agent-pantheon-implementation.md § Wave 7
ActionType schema each workflow consumes	action-ontology.md
Risk classification each decision resolves against	risk-classification.md