stag-techops

Plugin stag-techops from marketplace stag-skills in mridulaEA/stag. Commands read from the repo's manifest — 19 days ago.

Skills (21)

• stag-ai-automation-governance

  Govern production automations (runbook bots, agents, scheduled remediations) through STAG — detect failure spikes, retry storms, unknown outcomes and cost runaways per automation; pause safely as one approved operation; reconcile unknown outcomes from target state; measure cost per correct completed

• stag-crew-authoring

  Create, validate and run a STAG crew — a declarative composition of roles (agents + skills + permissions) and a workflow of backbone steps (coordinate, transition, investigate, propose_action, authorize, dispatch, verify, notify, record) in crews/<name>/crew.yaml with a scenario for end-to-end check

• stag-customer-care-comms

  Handle a wave of customer impact through STAG — detect ticket surges, CSAT drops and blocked journeys per product area and region, correlate duplicates into one episode, link to the TechOps incident, publish proactive and resolution communications once with provenance, apply an approved known-issue

• stag-dataops-pipeline-triage

  Triage and remediate data incidents with STAG — pipeline failures, data-quality drift, freshness SLA breaches and row-count collapses correlated per dataset; lineage-aware bounded investigation; idempotent backfill or quarantine as one approved operation; freshness and quality verification; single c

• stag-log-analysis

  Analyse log streams with STAG as evidence, not prompts — detect error signatures, volume bursts and collector lag; aggregate repeated lines into templates, counts and representative examples; run scoped read-only queries; prepare TTL-bounded sampling adjustments that preserve error levels; and assem

• stag-multi-agent-orchestration

  Design or run STAG's multi-agent incident workflow — coordinator, investigator, remediation reviewer and recovery verifier — with handoff schemas, separate permissions, termination rules and conflict resolution, built on the deterministic backbone (incident state machine, assignment contracts, actio

• stag-problem-management

  Run ITIL-style problem management on STAG incident history — detect recurring incidents and repeated workarounds, open a problem record that links (not merges) incident episodes, drive evidence-based root-cause analysis, draft a known-error record with workaround, and take a permanent fix through ch

• stag-release-gate

  Gate releases and changes with live STAG evidence — change-freeze windows, open incidents on the service and its dependencies, canary health, emergency/hotfix flags — and prepare a policy-bounded rollback as one retry-safe operation with a recorded release decision. Use when deployment or canary sig

• stag-security-governance

  Run enterprise AI and data governance on STAG's security streams — shadow AI usage, sensitive data to unsanctioned tools, acceptable-use violations, anomalous usage from DLP, proxy, identity and CASB — as deduplicated cases per tool and department with bounded exposure investigation, an approved DLP

• stag-skill-authoring

  Create, validate and contribute a STAG skill that works in Claude Code, Codex and Cursor from one SKILL.md source. Use when adding a new TechOps workflow, a role skill, or a development skill to the STAG skill pack, or when reviewing a community skill PR. Covers the Agent Skills frontmatter rules, f

• stag-stream-adapter

  Build a new STAG stream source / adapter (stag/streams) as a governed product — stable event identity, checkpointing, backfill and replay rules, schema-drift handling, fixtures and tests. Use when connecting Kafka topics, webhooks, SaaS audit logs, OpenTelemetry or any new source to STAG. Ensures re

• stag-techops-bounded-investigation

  Run or author a bounded, evidence-grounded incident investigation for STAG TechOps under an explicit assignment contract (scope, permitted capabilities, call/time limits, required output). Use when an InvestigationRequest exists, when asked "why is this happening", or when writing the investigate st

• stag-techops-dedup-correlation

  Decide what is actually repeating in a STAG stream and apply the right control — redelivery vs repeated symptom vs continuing signal vs duplicate ticket vs reingested document vs retried tool call. Use when counts look inflated, tickets multiply, "the same alert keeps firing", or when designing iden

• stag-techops-enterprise-onboarding

  Onboard an application, service or SaaS platform into enterprise STAG — register identity and ownership, define the normalization envelope, classify data (sensitivity, type, residency, trust, importance, purpose), compute eligible processing routes, and assign budgets. Use when connecting a new sour

• stag-techops-incident-triage

  Triage a new STAG signal or incident episode for TechOps. Use when a trigger fires, an alert pages, a new incident appears in /api/techops/incidents, or someone asks "is this real, how bad is it, who owns it?". Produces a triage record (identity check, impact, severity, ownership, paging decision) w

• stag-techops-postmortem

  Draft a blameless postmortem for a STAG incident from its recorded history (observations, transitions, investigations, actions, verification) and compute the inspectable metrics (false merges/splits, duplicate side effects, time to useful triage, verified restoration time, fully loaded cost). Use af

• stag-techops-recovery-verification

  Verify that a STAG incident has actually recovered after mitigation, using independent health signals over a defined observation window with hysteresis. Use when an action was dispatched, when someone says "looks fixed, can we close it?", or when deciding RESOLVED vs INVESTIGATING again. Treats miss

• stag-techops-remediation-ledger

  Propose, authorize, dispatch, reconcile and verify a remediation action for a STAG incident using durable operation identity so nothing is executed twice. Use when an investigation proposes a mitigation, when an operator asks to "restart/rollback/scale X", when a tool call timed out with unknown out

• stag-techops-resilience-tests

  Generate or review the failure-scenario test suite for a STAG deployment — redelivery, similar-but-distinct errors, false merges, concurrent workers, late events, reingested runbooks, lost action responses, LLM outage, prompt injection via telemetry, collector silence and historical replay. Use when

• stag-techops-signal-tuning

  Tune STAG's detection cadence and alert hygiene — debounce, cooldown, hysteresis, inhibition, burn-rate/multi-window SLO alerting, late-event and replay policy, backpressure. Use when alerts are noisy or flapping, when investigations fire too often or too rarely, when paging depends on AI, or when s

• stag-trigger-rule-authoring

  Add or modify a STAG trigger rule (stag/triggers/rules.py) with tests and registry wiring. Use when asked to detect a new condition in the stream, change thresholds, or when a rule fires too often or too rarely. Keeps rules deterministic, confidence-scored and covered by tests; defers cadence to the

Subagents (11)

• stag-automation-engineer

  Automation governance engineer for STAG. Use when runbook bots, agents or scheduled remediations show failure spikes, retry storms, unknown outcomes or cost runaways; prepares a safe pause as one ledger operation, coordinates bounded investigation, and drafts automation reviews (candidates, guardrai

• stag-customer-communicator

  Customer care communicator for STAG. Use when ticket surges, CSAT drops or blocked customer journeys form an episode; drafts proactive status updates and resolution follow-ups for the whole wave, prepares the known-issue response as one approved operation, and keeps customer text as untrusted input.

• stag-data-steward

  Data steward for STAG DataOps. Use when pipeline failures, data-quality drift, freshness SLA breaches or row-count collapses arrive; correlates per dataset, coordinates lineage investigation, prepares an idempotent backfill or quarantine as one approved operation, verifies freshness and quality, and

• stag-incident-coordinator

  Deterministic-first incident coordinator for STAG TechOps. Use to dedup and correlate fired triggers into incident episodes, decide whether a change is material, create investigation requests, and record lifecycle transitions through the API. Never calls the LLM path for repeated evidence and never

• stag-investigator

  Bounded, read-only incident investigator for STAG TechOps. Use when an investigation request exists and needs evidence-grounded analysis under an assignment contract (scope, permitted capabilities, call and time limits). Returns observed facts with references, hypotheses, contradictory or missing ev

• stag-log-analyst

  Log analyst for STAG. Use when error signatures, log-volume bursts or collector lag appear; turns repeated lines into templates, counts and representative examples, runs scoped read-only queries, prepares TTL-bounded sampling adjustments that never drop error levels, and assembles evidence packages

• stag-problem-analyst

  Problem-management analyst for STAG. Use to turn recurring incidents and repeated workarounds into a problem record, coordinate root-cause analysis, draft a known-error record and drive a permanent fix through change approval. Links incidents; never merges them away. Drafts stay drafts until reviewe

• stag-recovery-verifier

  Independent recovery verifier for STAG TechOps. Use after a remediation was dispatched or when someone wants to close an incident. Checks externally observable health and customer impact over a defined window with hysteresis, treats missing telemetry as unknown, and records the verdict on the action

• stag-release-gatekeeper

  Release gatekeeper for STAG. Use when deployment and canary signals arrive to check change freezes, open incidents on the service and its dependencies, and canary health; prepares the release decision and, when needed, a rollback proposal bounded by policy. Read-only; never deploys or rolls back by

• stag-remediation-reviewer

  Turns an investigation's proposal into one narrowly scoped, retry-safe remediation operation in STAG's action ledger and prepares the evidence an approver needs. Use when findings propose a mitigation, when an operator asks to restart, roll back or scale something, or when a dispatched action has an

• stag-security-analyst

  Security and AI-governance analyst for STAG. Use when DLP, proxy, identity or CASB streams show sensitive data flowing to unsanctioned AI tools, acceptable-use violations or anomalous usage; correlates per tool and department, coordinates bounded exposure investigation, prepares a DLP control as one

Similar plugins