AXAxiom Agent · deployment capabilities
A focused inventory of the release-governance logic behind Axiom: assess a change, prepare evidence, require human approval, and verify the result. A listed capability proves application logic exists; it does not by itself claim a live integration or automated deployment.
One governed Playbook
A capability is not a promise of autonomous execution or a live integration. It is a governed part of the workflow.
01
Bring typed request, repository, environment, and signal context into a record people can review.
02
Keep intended steps, checks, recovery context, and evidence requirements close to the decision.
03
Make policy limits and human approval requirements visible before an action is considered.
04
Separate planning and recorded handoff from a verified provider action or production result.
05
Preserve validation, recovery context, versions, and attributable evidence for later review.
These details are available for technical review. They do not claim live cloud mutation, a configured integration, or autonomous deployment.
The operation-planning kernel batches resources to the configured maximum, rejects an invalid oversized step, and records a boundary event. Example: a 12-resource plan with a five-resource limit becomes 5 + 5 + 2 planning batches. This is kernel enforcement; live cloud mutation enforcement is not release-verified.
The application-side module stores action outcome metadata in the audit store under an organization and cloud-account context. A previous failure can downgrade guidance for that resource to human review. Example: a failed resize recommendation makes the next resize proposal more cautious. Consistent retention/deletion controls and the released desktop journey remain unverified.
The underlying controls remain available for engineering review, without turning this page into a wall of implementation cards.
Detector
Watches telemetry and emits typed signals to the bus.
lib/agents/detectorSignalEmitter
→ agent busReasoner
Weaves 1..N signals into a typed hypothesis with a confidence score.
lib/agents/reasonerHypothesisWeaver
→ agent busSimulator
Models a proposed change against constrained inputs and records review checks before an approval packet is prepared; it is not an isolated production rehearsal.
lib/agents/simulatorSandboxSpec
Spec Writer
Drafts a typed engineering spec (goals, non-goals, risks, verification, rollback) from a problem statement.
lib/agents/specWriter
Test Coverage Proposer
Reads a diff and proposes the must-have, should-have, and nice-to-have tests with rationale per row.
lib/agents/testCoverageProposer
Refactor Sequencer
Orders a multi-step refactor with per-step verdict (safe / needs-test / needs-review / blocked) and rollback.
lib/agents/refactorSequencer
Migration Coordinator
Builds a multi-stage migration runbook with backwards-compat window, gate checks, and rollback per stage.
lib/agents/migrationCoordinator
GitHub Pipeline Repairer
Classifies a failed Actions run and proposes a typed PR patch (retry, pin, quarantine, raise runner, etc.).
lib/agents/githubPipelineRepairer
Release Notes Drafter
Drafts release notes from commits between two tags. Conventional-commit aware, recommends SemVer bump.
lib/agents/githubReleaseNotesDrafter
Incident Timeline Weaver
Weaves heterogeneous events into a phase-tagged incident timeline with TTD / TTM / TTR durations.
lib/agents/incidentTimelineWeaver
Postmortem Drafter
Renders a ready-to-merge postmortem from a timeline + action items. Refuses to draft while incident is ongoing.
lib/agents/postmortemDrafter
Database Schema Reviewer
Audits tables / columns / indexes / FK constraints. Surfaces missing-FK, missing-index, naming drift, duplicate indexes, orphan columns.
lib/agents/databaseSchemaReviewer
Slow Query Proposer
Reads slow-query rows + schema, proposes add_index (with DDL), rewrite_query, partition_table, downsize, or no_action.
lib/agents/slowQueryProposer
Alert Noise Reducer
Analyzes alert fire / ack / incident rates, proposes raise_threshold / widen_window / require_paired_metric / mute / split_routing.
lib/agents/alertNoiseReducer
Workflow Orchestrator
Chains kernels into typed workflows with per-step gate handoff. Closed-union step states; safety-tier rejection blocks downstream dispatch.
lib/agents/agentWorkflowOrchestrator
Intent Parser
Natural-language operator intent → typed ActionSpec (kind + domain + urgency + entities + destructive flag).
lib/agents/intentParser
Anomaly Detector
MAD-based z-score on time series. Spike / dip / drift / missing_data with bounded false-positive rate.
lib/agents/anomalyDetector
Change Risk Assessor
Scores blast radius of a proposed change against service topology. Critical-tier reach → council_3_of_5 gate.
lib/agents/changeRiskAssessor
Secrets Hygiene
Scans a file / diff for AWS / Stripe / GitHub / OpenAI / Anthropic / PEM / Slack secrets. Redacts in output; never echoes the raw value.
lib/agents/secretsHygieneScanner
Compliance Mapper
Maps available audit events to control-review evidence and highlights gaps. It supports customer assessment; it does not certify compliance or regulated-workload readiness.
lib/agents/complianceControlMapper
Policy Gate
Applies the tenant charter to every proposal — refuses anything outside the operator-signed scope.
lib/agents/policyGateEvaluator
→ autonomy charterBoundary Gate
Classifies the blast radius of every proposal into a closed-union severity tier.
lib/agents/boundaryGateCatalog
→ boundary catalogCouncil
Weighted-vote consensus — ⅔ default — across the planning + safety agents.
lib/agents/council
→ agent busApprover
Assembles the approval context an operator reviews. A recorded approval does not itself execute a provider action.
lib/agents/approverPacketAssembler
→ approvalsVerifier
Records configured validation outcomes for review. It does not claim universal provider verification or production health.
lib/agents/verifierPostExecChecker
→ audit logAuditor
Records a rationale digest and decision context to support later review of the governed workflow.
lib/agents/auditorRationaleWriter
→ audit logImprover
Proposes method improvements from the audit trail — the loop that turns evidence into a better proposal.
lib/agents/improverProposalSynthesizer
→ method proposalsConfidence Calibrator
Wilson-lower-bound calibration of kernel accuracy from approval / reject / rollback outcomes. Recommends trust_more / pause / trust_less.
lib/agents/confidenceCalibrator
Activity Aggregator
Aggregates per-agent activity into the operator-visible cockpit timeline.
lib/agents/agentActivityAggregator
→ agent activityMemory Consolidator
Compresses raw observations into episodic / semantic / procedural records. Recurrent observations promote to semantic. Prunes by relevance + age.
lib/agents/agentMemoryConsolidator
Provider configuration, authorization, persistence, recovery, and a verified customer journey must also be present before a capability is described as available.