DEV Community: edgehawk

Cisco command collection: keep sending, execution, and log verification separate

edgehawk — Mon, 07 Sep 2026 22:28:03 +0000

I develop FlowChain, a Windows utility for sending an ordered list of text items to another application. This post introduces a practical runbook workflow and a new browser demo. The demo is simulated; it is not a physical Cisco-device test.

Three different things to verify

When collecting network evidence, a highlighted row in a helper app answers only one question: did the helper send that item? It does not prove that the command succeeded or that the session log contains all of its output. Keeping these checks separate makes the workflow easier to review:

Sending: which planned command did I send?
Execution: did the intended device accept it, and did I read the complete response?
Collection: does the saved log contain the planned commands in the expected order?

A small collection list

show running-config
show ip route
show ip interface brief

These are examples for inspecting configuration, routes, and interface state. Verify support, privilege requirements, and the collection scope for your particular platform. A command list is not a universal maintenance procedure.

An operator-controlled workflow

Connect to the intended device using Tera Term or PuTTY. Check the hostname and prompt, and start session logging under your normal operating procedure.

Capture the reviewed command list in a dedicated FlowChain tab and select the terminal as the destination. Send one line using the global hotkey. Check whether automatic Enter is enabled, inspect the result, and wait for the prompt before sending the next item.

If a long response stops at --More--, finish reading the paginated output first. A visible prompt, an error message, and a partial log all mean different things. Do not assume that advancing the helper list validates the device output.

Compare the log afterward

FlowChain can compare sendable rows in the current tab with a selected Tera Term or PuTTY log. The results distinguish collected, out-of-order, and missing commands. Selecting a result takes you to a relevant position in the log for investigation. Skip-marked rows are excluded.

That check is about command presence and order. It is not a guarantee of successful execution, complete output, or a correct network configuration. Review those separately. Logs can contain sensitive configuration information, so inspect them before sharing. The selected log is processed locally on the PC.

Try the workflow

The new Cisco collection guide includes a browser simulation: click Send and see sample responses for the three commands above. No device is connected.

There is also a 30-second runbook overview. It demonstrates the input workflow, rather than a physical Cisco lab.

The free edition provides basic one-item sending and two tabs. A one-time Pro purchase adds capabilities such as more tabs and continuous sending. FlowChain on Microsoft Store.

If you collect command output during maintenance, which step causes more rework: locating the next command, handling pagination, or checking the saved log?

Reference: Cisco IOS XE IP Unicast Routing guide.

Disclosure: I am the developer of FlowChain. This article was prepared with AI assistance and checked against the current product implementation.

A Safer Way to Run Multi-Vendor Infrastructure CLI Runbooks

edgehawk — Sun, 30 Aug 2026 03:59:02 +0000

Infrastructure work rarely stays inside one vendor ecosystem.

A single maintenance window may include an out-of-band management controller, a storage array, several network switches, a firewall, a hypervisor, and a UPS or PDU. The interfaces differ, but the operator's problem is often the same: move through a reviewed sequence of commands without losing the current console, skipping a step, or pasting too much at once.

This article describes a simple operator-controlled workflow for multi-vendor runbooks on Windows.

The real problem is context switching

Imagine a change that touches:

Dell PowerEdge iDRAC or RACADM
HPE ProLiant iLO
IBM Power HMC or Lenovo XClarity
NetApp ONTAP or Pure Storage Purity
Cisco IOS XE or NX-OS
Yamaha RTX, NVR, or SWX
Juniper Junos, Fortinet FortiOS, or Palo Alto PAN-OS
VMware shells, serial servers, KVM devices, UPS units, and PDUs

The commands are not interchangeable. Authentication models, prompts, configuration modes, commit behavior, and rollback procedures vary. Yet many teams still copy instructions from a document, return to the console, paste, check the result, and repeat.

That repeated window switching creates avoidable risk:

The wrong line can be copied.
A command can be skipped or repeated.
A multi-line paste can run faster than the operator can review.
Focus can move to another terminal without being noticed.
Output can be mistaken for the next input step.

The goal should not be unattended automation. The goal should be a controlled, reviewable input sequence.

Use one item per operational step

A safer runbook format keeps one deliberate action in each item.

For example:

show version
show inventory
show interfaces status
show logging

The actual syntax must come from the device's official documentation and your approved change procedure. The important part is the structure: one step, one review point, one operator decision.

This format works for more than network switches. The same pattern can organize:

server hardware inventory and health checks;
out-of-band controller diagnostics;
storage capacity, volume, controller, and replication checks;
firewall policy inspection;
hypervisor host validation;
UPS and PDU status collection.

Keep explanatory notes outside the executable item whenever possible. If a note must sit beside a command, make the distinction visually obvious before the maintenance window.

Keep the operator in control

A useful sequential-input tool should reduce mechanical work without making operational decisions.

My preferred sequence is:

Load the reviewed runbook.
Put each command or value in a separate item.
Open the correct console.
Confirm the hostname, device role, environment, and current privilege level.
Send exactly one item.
Read the output.
Decide whether the next item is still appropriate.
Stop immediately if the result differs from the expected state.

This is intentionally slower than blind bulk execution. It is usually faster than manual copy-and-paste while preserving a human checkpoint between commands.

Separate read-only validation from change commands

Do not mix every action into one undifferentiated list.

A practical structure is:

Phase 1: identity and baseline

Confirm the device, firmware or software version, hardware inventory, active controller, cluster role, and current alarms.

Phase 2: pre-change validation

Check interfaces, sessions, replication, failover state, configuration status, and available capacity.

Phase 3: controlled change

Use the approved vendor-specific commands. Insert explicit pause points before destructive or disruptive operations.

Phase 4: post-change validation

Repeat the relevant baseline checks and compare results with the expected state.

Phase 5: evidence capture

Save terminal output through your organization's approved logging process. A sequential-input tool should not be treated as the system of record.

This phased approach applies across server, storage, and network platforms even though the commands themselves differ.

Guardrails that matter in real environments

Before sending anything, verify:

the focused window and tab;
the device hostname and management IP;
production versus staging;
the current user and privilege level;
whether the console supports bracketed paste or has unusual input handling;
line ending behavior;
whether Enter is sent automatically;
whether the command expects confirmation;
whether the operation has a rollback or recovery path.

For high-risk changes, use a second-person review and a documented maintenance plan. Do not store passwords, private keys, tokens, or recovery codes in a reusable command list.

What FlowChain contributes

FlowChain is a Windows input assistant that keeps a list of text items and sends them one at a time with a button or hotkey. That lets an operator remain in an RDP, SSH, serial, web-console, PowerShell, or SQL window while advancing through a reviewed sequence.

The useful part is not vendor-specific automation. It is the interaction model:

one item at a time;
explicit operator action;
visible progress through the list;
fewer trips back to the source document;
compatibility with consoles where ordinary clipboard paste is inconvenient.

Vendor and product names are workflow examples, not certified integrations or compatibility claims. Every environment must be tested safely for focus behavior, keyboard layout, paste handling, syntax, permissions, and device-specific confirmation prompts.

A small runbook design checklist

Before the maintenance window:

validate every command against the correct model and software release;
remove credentials and secrets;
split compound actions into reviewable steps;
mark expected results and stop conditions;
identify commands that change state;
rehearse in a lab or non-production environment;
confirm backup, rollback, and escalation procedures.

During the window:

verify the target before every state-changing section;
advance only after reading the previous result;
stop on unexpected prompts or output;
keep your change record separate from the input list.

After the window:

run the post-change checks;
preserve evidence according to policy;
update the runbook with verified lessons;
remove temporary operational data.

Try the multi-vendor workflow guide

I published a more detailed guide covering server management, storage, network and security consoles, plus other infrastructure devices:

IT Infrastructure CLI Runbooks for Servers, Storage and Networks

The principle is simple: reduce repetitive window switching, but keep every operational decision with the engineer.

Paste Line by Line into RDP, SSH, SQL, and Runbooks without Switching Windows

edgehawk — Tue, 28 Jul 2026 12:00:43 +0000

When working through a runbook, pasting the whole command block is often the wrong move.

Sometimes you need to check each command first. Sometimes you need to wait for the previous step to finish. Sometimes one destructive line in the wrong window is enough to ruin the afternoon.

I built FlowChain for this exact kind of careful input work on Windows.

FlowChain captures a list once, then sends the next item to another app with a global hotkey. You can keep your destination window focused and move through the list one item at a time.

Official site and demo:

https://flowchain-official.blueedge14.chatgpt.site

The Problem

A lot of Windows work still involves moving ordered text from one place to another:

spreadsheet rows into web forms
canned replies into support tools
QA values into test fields
SSH or SQL commands into a terminal
RACADM or network-device steps into a remote session
AI-generated command blocks into a shell after review

Normal copy and paste is fast, but it is also coarse. You copy a block, switch windows, paste, switch back, copy the next line, switch again, and repeat.

Or worse, you paste the whole block and hope the target application handles it the way you expected.

What FlowChain Does

FlowChain turns a block of text or table data into an ordered send list.

The basic loop is:

Capture or paste a list into FlowChain.
Keep the destination app active.
Press a global hotkey to send the next item.
FlowChain marks completed rows so you can resume later.

It can send through normal clipboard paste, or it can type characters as real keystrokes. That helps in places where normal paste is unreliable, such as RDP, VNC, virtual machines, terminal sessions, and older applications.

Useful Cases

FlowChain is not trying to replace full automation tools. It is for the middle ground where you still want human review and control.

Good fits include:

entering rows from Excel into business systems
sending support replies one message at a time
registering products, SKUs, and descriptions
feeding QA test data into forms
stepping through RDP or VM procedures
running SSH, SQL, RACADM, or network runbooks line by line
reviewing AI-generated commands before sending each one

Install

Microsoft Store:

https://apps.microsoft.com/detail/9NDSWGBHLV2J

winget:

winget install --id 9NDSWGBHLV2J --source msstore

The free edition covers the core line-by-line workflow. The optional Pro upgrade removes ads and unlocks higher limits and additional automation features.

Disclosure: I am the developer of FlowChain.