IT OPS
Agent-Driven Cloudflare Access Drift Investigator
A CEO-driven agent reconciles Cloudflare Access grants against the HR system over a custom MCP server, investigates each discrepancy with context.
How it runs
The automated pipeline, trigger to output.
- TriggerScheduled agent run (twice weekly)
- ActionEnumerate Cloudflare Access apps and grantsCloudflare
- ActionQuery HR system of record via custom MCPCustom MCP server
- LogicAgent reasons per grant; assigns keep/revoke/escalate
- ActionPersist dispositions and evidence to PostgresPostgres
- OutputPost prioritized findings to Slack for approvalSlack
What it does
Uses an agent rather than a fixed pipeline to reason about access drift. It compares Cloudflare Access grants to the HR system of record exposed through your custom MCP server, then investigates each mismatch, distinguishing genuine orphans from contractors, service accounts, and recently-rehired staff before recommending keep, revoke, or escalate.
When to use it
Use this when naive set-diffing produces too many false positives and you want judgment applied: shared mailboxes, vendor identities, and edge cases that a deterministic rule mislabels. Best when access ownership data is messy.
How it works
- 1A scheduled run kicks off the agent.
- 2The agent calls Cloudflare to enumerate Access apps and their grants.
- 3It queries the HR system of record through the custom MCP server for each identity's employment context.
- 4For each grant lacking a clean active match, the agent reasons over role, contractor flags, and rehire dates to assign a disposition with a rationale.
- 5It writes its dispositions and evidence to Postgres.
- 6It posts a prioritized findings summary to Slack for human approval.
Set it up
What you configure once, before turning it on.
- 1Connect CloudflareWorkers, Pages, R2, KV — the edge stack.
- 2Connect Custom MCP serverConnect any MCP-compatible tool you own.
- 3Connect PostgresAny Postgres URL — query, write, migrate.
- 4Connect SlackChannels, DMs, threads, mentions.
- 5Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
- 6Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
- 7Test, then turn it onRun once against a sample, confirm the output, then enable the trigger.
More IT Ops workflows
Daily Building Anomaly Digest to MS Teams
Each morning queries BigQuery for the prior day's flagged sensor anomalies, summarizes them by site and system into a ranked briefing.
Indoor Air Quality Breach to Tenant Notice and Work Order
Listens for CO2, VOC, or humidity sensor alerts via webhook, and when a zone exceeds occupant-safety limits it emails affected tenants, opens a Monday remediation task.
Self-Service Reclaim Email for Idle Users
Detects users idle in a SaaS app past the threshold and emails each one a keep-or-release link; unanswered seats after the deadline are auto-flagged for removal.
Outlook Room Conflict Resolver with Approval Gate in Teams
When an Outlook room clashes, proposes a rebooking and asks the bumped meeting's organizer to approve the move in Microsoft Teams before any change is made.
Outlook Room Double-Booking Resolver with Auto-Rebook
Detects when two meetings claim the same Outlook room resource and automatically relocates the lower-priority meeting to a comparable free room.
Monthly Wasted-License Cost Report
Aggregates inactive-seat data across all tracked SaaS apps each month, computes total reclaimable spend, and delivers a ranked cost report to leadership in Notion and Slack.
Run it inside a business
This workflow drops into a full company template. Import the org, and this is one of the playbooks its agents run.
Run this workflow in your colony.
14-day trial. No DevOps. No Sales call. Provisioned in under a minute.