Binary reference

workflow-manager

Workflow manager for batch command execution with conditional logic

Version9.8.4 (build 320)Size8.6MBAuthorWarith Al MaawaliLicenseProprietaryWebsitehttps://www.digi77.com

File Information

Property	Value
Binary Name	workflow-manager
Version	9.8.4 (build 320)
Build Date	Not disclosed
Rust Version	1.82.0
File Size	8.6MB
Author	Warith Al Maawali
License	Proprietary
Category	Kodachi Binary
Description	Workflow manager for batch command execution with conditional logic
Git Commit	unknown
Metadata Generated	2026-08-10T13:23:00Z
Binary Timestamp	Unknown
JSON Data	View Raw JSON

SHA256 Checksum

c14889d46b98fef8554e0b466e8ed06ebabca812864ea76f07ba9257a87be2a5

Features

#	Feature
1	Template-based workflow management
2	Conditional command execution
3	Batch processing with retry logic
4	State tracking and logging
5	Concurrent execution support
6	Pause steps with user confirmation
7	Substring, regex and JSON path conditions
8	JSON path subset support (dot fields + array indexes)
9	Prerequisites validation before execution
10	System state checking and probes
11	Reusable probe functions for conditions

Security Features

Feature	Description
Authentication	Not provided by cli-core (see online-auth)
Encryption	Not provided by cli-core
Input Validation	Argument parsing via clap; per-command validation is the consumer's responsibility
Rate Limiting	Not provided by cli-core

System Requirements

Requirement	Value
OS	Linux (Debian-based)
Privileges	root/sudo for system operations
Dependencies	OpenSSL, libcurl

Global Options

Flag	Description
`-h, --help`	Print help information
`-v, --version`	Print version information
`-n, --info`	Display detailed information
`-e, --examples`	Show usage examples
`--json`	Output in JSON format
`-o, --output-format <FORMAT>`	Force output format (text\|json)
`--json-pretty`	Pretty-print JSON output with indentation
`--json-human`	Enhanced JSON output with improved formatting (like jq)
`--json-filter <FIELD1,FIELD2>`	Return only specified JSON fields (comma-separated)
`--fields <FIELD_LIST>`	Select specific fields to include in output (comma-separated)
`--limit <NUMBER>`	Limit number of results returned
`--offset <NUMBER>`	Skip first N results (for pagination)
`--verbose`	Enable verbose output
`--quiet`	Suppress non-essential output
`--no-color`	Disable colored output

Commands

Diagnostics

`prereq`

Validate workflow prerequisites

Usage:

workflow-manager prereq <check|validate> <TEMPLATE>

Options:

<SUBCOMMAND>: check, or validate (an alias for check)
<TEMPLATE>: Template whose prerequisites block is evaluated

Examples:

workflow-manager prereq check initial_terminal_setup_wireguard_torrify

workflow-manager prereq validate my-workflow

workflow-manager prereq check my-workflow --json

`state`

Query system state

Usage:

workflow-manager state [FIELD]

Options:

[FIELD]: Single state to query (online, torrify, tor_running, security_score, routing_mode, authenticated, ...). Omit to return all 20
--json: Emit the state as a JSON envelope (global flag)

Examples:

workflow-manager state

workflow-manager state online

workflow-manager state security_score --json

Execution

`run`

Run a workflow template

Usage:

workflow-manager run <TEMPLATE> [OPTIONS]

Options:

<TEMPLATE>: Template to execute
--dry-run: List the steps that would run without executing any of them
--kill-policy <POLICY>: What to do when a step fails: stop, continue or skip_remaining. Overrides global_settings.kill_policy
--max-log-size <BYTES>: Cap the run log size in bytes. Overrides global_settings.max_log_size
--param <KEY=VALUE>: Override a declared workflow parameter. Repeat for multiple parameters
-y, --yes: Skip ordinary top-level [y/N] prompts in text mode; never approves high-risk or included child steps
--param <NAME=VALUE>: Override a workflow parameter; may be repeated
--approved-step <JSON>: Carry one exact dashboard-approved child step record; may be repeated

Examples:

workflow-manager run my-workflow --dry-run

workflow-manager run my-workflow

workflow-manager run my-workflow --kill-policy continue

workflow-manager run my-workflow --max-log-size 20971520

workflow-manager run my-workflow --param interface=eth0 --param mode=safe

workflow-manager run my-workflow --yes

workflow-manager run my-workflow --json

Steps

`add`

Add a step to a workflow template

Usage:

workflow-manager add <TEMPLATE> <COMMAND> [OPTIONS]

Options:

<TEMPLATE>: Template to append the step to
<COMMAND>: Command to execute. A comma separated list creates one step per command
--description <TEXT>: Step description
-c, --condition <TYPE>: Basic condition: always, if_success or if_fail
--if-probe <NAME>: Run only if the named probe from the profile's probes block matches
--if-probe-expect <JSON>: Value the probe must return (JSON: true, 4, "ok")
--if-state <NAME>: Run only if a live system state matches (see `workflow-manager state`)
--if-state-expect <JSON>: Value the state must have (JSON: true, false, "wireguard")
--if-expression <EXPR>: Boolean expression over probe('NAME') and state.NAME, joined with && and ||
--if-contains <TEXT>: Run only if the previous step's output contains TEXT (substring, case sensitive)
--if-not-contains <TEXT>: Run only if the previous output does not contain TEXT
--if-equals <VALUE>: Run only if the previous output equals VALUE exactly (trimmed)
--if-regex <PATTERN>: Run only if the previous output matches the regex
--if-not-regex <PATTERN>: Run only if the previous output does not match the regex
--if-json-path <EXPR>: Run only if a JSON field matches, e.g. '$.status="connected"' (writes if_json_path_equals)
-t, --timeout <SECONDS>: Step timeout. Falls back to global_settings.default_timeout when omitted
--confirm: Prompt [y/N] before this step runs (text mode only)
-w, --working-dir <PATH>: Directory to run the step from
--nice-level <N>: Process priority, -20 (highest) to 19 (lowest). Negative values need root
--env <VARS>: Comma separated allowlist of environment variables passed to the step

Examples:

workflow-manager add my-workflow 'echo Hello World'

workflow-manager add my-workflow 'tar czf backup.tar.gz /data' -t 600 --description 'Archive data'

workflow-manager add my-workflow 'cleanup.sh' -c if_success

workflow-manager add my-workflow './notify.sh' --if-contains 'ERROR'

workflow-manager add my-workflow './handle-connected.sh' --if-json-path '$.status="connected"'

workflow-manager add my-workflow './tor-ready.sh' --if-state tor_running --if-state-expect true

workflow-manager add my-workflow './low-score-fix.sh' --if-expression "state.security_score < 60"

workflow-manager add w1 "ip-fetch","online-auth check-login","ip addr show"

`pause`

Add a pause step to a workflow template

Usage:

workflow-manager pause <TEMPLATE> [OPTIONS]

Options:

<TEMPLATE>: Template to append the pause to
-m, --message <TEXT>: Message shown to the user while the workflow waits
-c, --condition <TYPE>: Basic condition: always, if_success or if_fail

Examples:

workflow-manager pause my-workflow -m 'Review results before continuing'

workflow-manager pause backup -m 'Verify backup integrity' -c if_success

`include`

Add an include step to a workflow template (includes another profile)

Usage:

workflow-manager include <TEMPLATE> <PROFILE> [OPTIONS]

Options:

<TEMPLATE>: Template to append the include to
<PROFILE>: Existing profile whose steps are expanded in place at run time
--description <TEXT>: Step description
-c, --condition <TYPE>: Basic condition: always, if_success or if_fail
--if-probe <NAME>: Run only if the named probe from the profile's probes block matches
--if-probe-expect <JSON>: Value the probe must return (JSON: true, 4, "ok")
--if-state <NAME>: Run only if a live system state matches (see `workflow-manager state`)
--if-state-expect <JSON>: Value the state must have (JSON: true, false, "wireguard")
--if-expression <EXPR>: Boolean expression over probe('NAME') and state.NAME, joined with && and ||

Examples:

workflow-manager include my-workflow shared-prereqs --description 'Reusable setup steps'

workflow-manager include my-workflow tor-recovery -c if_fail

workflow-manager include my-workflow base-recovery-sequence --if-state online --if-state-expect true

`update`

Update a step in a workflow template

Usage:

workflow-manager update <TEMPLATE> <STEP_ID> [COMMAND] [OPTIONS]

Options:

<TEMPLATE>: Template to edit
<STEP_ID>: Step number to update, as shown by `workflow-manager show`
[COMMAND]: New command for a command step. Omit it to change only the flags below
-m, --message <TEXT>: New message for a pause step
--profile <NAME>: New profile for an include step
--description <TEXT>: New step description
-c, --condition <TYPE>: Basic condition: always, if_success or if_fail
--if-probe <NAME>: Run only if the named probe from the profile's probes block matches
--if-probe-expect <JSON>: Value the probe must return (JSON: true, 4, "ok")
--if-state <NAME>: Run only if a live system state matches (see `workflow-manager state`)
--if-state-expect <JSON>: Value the state must have (JSON: true, false, "wireguard")
--if-expression <EXPR>: Boolean expression over probe('NAME') and state.NAME, joined with && and ||
--if-contains <TEXT>: Run only if the previous step's output contains TEXT (substring, case sensitive)
--if-not-contains <TEXT>: Run only if the previous output does not contain TEXT
--if-equals <VALUE>: Run only if the previous output equals VALUE exactly (trimmed)
--if-regex <PATTERN>: Run only if the previous output matches the regex
--if-not-regex <PATTERN>: Run only if the previous output does not match the regex
--if-json-path <EXPR>: Run only if a JSON field matches, e.g. '$.status="connected"' (writes if_json_path_equals)
--clear-condition: Drop the stored condition payload and reset to the basic type given by -c
-t, --timeout <SECONDS>: New step timeout
--confirm <BOOL>: Set or clear the confirm flag (true or false)
-w, --working-dir <PATH>: New working directory for the step
--nice-level <N>: New process priority, -20 (highest) to 19 (lowest)
--env <VARS>: New comma separated environment allowlist

Examples:

workflow-manager update my-workflow 1 'echo Updated Command' -t 300

workflow-manager update my-workflow 2 --if-contains 'ERROR' --description 'Alert on error'

workflow-manager update my-workflow 2 --clear-condition -c always

workflow-manager update my-workflow 3 -m 'Verify the archive, then press Enter' --confirm true

workflow-manager update my-workflow 4 --profile base-recovery-sequence

`delete-step`

Delete a step from a workflow template

Usage:

workflow-manager delete-step <TEMPLATE> <STEP_ID>

Options:

<TEMPLATE>: Template to edit
<STEP_ID>: Step number to remove, as shown by `workflow-manager show`

Examples:

workflow-manager delete-step my-workflow 2

workflow-manager show my-workflow && workflow-manager delete-step my-workflow 3

Templates

`create`

Create a new workflow template

Usage:

workflow-manager create <NAME> [OPTIONS]

Options:

<NAME>: Name of the template, stored as config/profiles/<NAME>.json
--description <TEXT>: Human readable description of the workflow
--force: Overwrite an existing template with the same name

Examples:

workflow-manager create my-workflow

workflow-manager create backup-workflow --description 'Daily backup routine'

workflow-manager create diagnostics && workflow-manager add diagnostics './health-control net-check'

`list`

List all workflow templates

Usage:

workflow-manager list

Options:

--json: Emit the template list as a JSON envelope (global flag)

Examples:

workflow-manager list

workflow-manager list --json

`show`

Show details of a workflow template

Usage:

workflow-manager show <TEMPLATE>

Options:

<TEMPLATE>: Template to display
--json: Print the raw profile JSON, which is exactly what the runner deserializes (global flag)

Examples:

workflow-manager show my-workflow

workflow-manager show my-workflow --json

`delete`

Delete an entire workflow template

Usage:

workflow-manager delete <TEMPLATE>

Options:

<TEMPLATE>: Template to delete. This removes the profile JSON from disk and cannot be undone

Examples:

workflow-manager delete my-workflow

workflow-manager list && workflow-manager delete obsolete-workflow

Operational Scenarios

Scenario-oriented workflows generated from the binary's built-in -e --json examples.

Scenario 1: Quick Start

Create a template, add a step, preview it, then run it. Templates are stored as JSON profiles under the hooks config directory (config/profiles/<NAME>.json).

Step 1: Create an empty workflow template

workflow-manager create my-workflow

Expected Output: Template 'my-workflow' created successfully

Step 2: Append a command step to the template

workflow-manager add my-workflow 'echo Hello World'

Expected Output: Step 1 added to template 'my-workflow'

Note

`add` requires the template to exist already, so run `create` first.

Step 3: Inspect the steps, their IDs and their conditions

workflow-manager show my-workflow

Expected Output: Workflow Template: my-workflow Step 1: echo Hello World (condition: always)

Step 4: Preview execution without running any command

workflow-manager run my-workflow --dry-run

Expected Output: Dry run: 1 step would be executed

Note

Always dry-run first: steps run real shell commands.

Step 5: Execute every step in order

workflow-manager run my-workflow

Expected Output: Workflow completed: Success (1/1 steps)

Note

workflow-manager itself does not need root. Individual steps may, if the command inside the step does.

Scenario 2: Template Management

Create, list, view, and delete workflow templates.

Step 1: Create a template with a human readable label

workflow-manager create backup-workflow --description 'Daily backup routine'

Expected Output: Template 'backup-workflow' created successfully

Step 2: Show every stored workflow template

workflow-manager list

Expected Output: Workflow Templates (2 total) backup-workflow (4 steps) my-workflow (1 step)

Step 3: Print the raw profile JSON, which is what the runner deserializes

workflow-manager show backup-workflow --json

Expected Output: The full profile object: name, description, global_settings, prerequisites, probes and steps

Note

Use this to see the exact schema before hand editing config/profiles/backup-workflow.json.

Step 4: Remove an entire workflow template

workflow-manager delete backup-workflow

Expected Output: Template 'backup-workflow' deleted successfully

Note

Destructive and irreversible: the profile JSON is deleted from disk. Use `delete-step` if you only want to remove one step.

Scenario 3: Adding Steps

Add command steps to a template. A comma separated command list creates one step per command.

Step 1: Add a step with a 10 minute timeout and a description

workflow-manager add my-workflow 'tar czf backup.tar.gz /data' --timeout 600 --description 'Archive data'

Expected Output: Step 2 added to template 'my-workflow' (timeout 600s)

Note

--timeout (-t) is the STEP timeout. It shadows the cli-core global --timeout inside `add`.

Step 2: Create a template and add three steps in one call using comma separation

workflow-manager create w1 && workflow-manager add w1 "ip-fetch","online-auth check-login","ip addr show"

Expected Output: Template 'w1' created, 3 steps added to template 'w1'

Note

All steps created in one call share the same timeout and condition settings.

Step 3: Run this step only if the previous step exited 0

workflow-manager add my-workflow 'cleanup.sh' -c if_success

Expected Output: Step added with condition if_success

Step 4: Mark a step as requiring an interactive confirmation before it runs

workflow-manager add my-workflow './health-control panic hard' --confirm

Expected Output: Step added with confirm=true

Note

confirm=true prompts [y/N] before the step runs, in TEXT mode only. A --json run (what the dashboard uses) proceeds past it without prompting, so confirm is a human speed bump, not a safety gate.

Step 5: Run a background friendly step from a specific working directory

workflow-manager add my-workflow './long-scan.sh' --nice-level 10 -w /opt/kodachi/dashboard/hooks

Expected Output: Step added with nice_level=10 and working_dir set

Note

nice-level range is -20 (highest priority) to 19 (lowest). Negative values need root.

Step 6: Allow only the listed environment variables through to the step

workflow-manager add my-workflow './report.sh' --env 'HOME,PATH,LANG'

Expected Output: Step added with env_allowlist [HOME, PATH, LANG]

Note

Comma separated. Anything not listed is stripped from the step environment.

Scenario 4: Conditional Logic (CLI flags)

Every condition the runner understands, expressed as an `add` flag. One condition per step.

Step 1: Always execute (this is the default)

workflow-manager add my-workflow './check-status.sh' -c always

Expected Output: Step added with condition always

Step 2: Execute only if the previous step exited 0

workflow-manager add my-workflow './deploy.sh' -c if_success

Expected Output: Step added with condition if_success

Note

The most common condition for sequential workflows.

Step 3: Execute only if the previous step exited non-zero

workflow-manager add my-workflow './rollback.sh' -c if_fail

Expected Output: Step added with condition if_fail

Note

Useful for error recovery and rollback.

Step 4: Execute if the previous step's output CONTAINS the text

workflow-manager add my-workflow './alert-success.sh' --if-contains 'success'

Expected Output: Step added with condition if_contains

Note

Plain substring match, case sensitive. No wildcards: use --if-regex if you need a pattern.

Step 5: Execute if the previous output does NOT contain the text

workflow-manager add my-workflow './continue.sh' --if-not-contains 'error'

Expected Output: Step added with condition if_not_contains

Note

Substring match, case sensitive.

Step 6: Execute if the previous output is exactly this value

workflow-manager add my-workflow './handle-done.sh' --if-equals 'ready'

Expected Output: Step added with condition if_equals

Note

Exact match, case sensitive. Output is trimmed before comparison.

Step 7: Execute if the previous output matches a regex

workflow-manager add my-workflow './process-result.sh' --if-regex '^status: (ok|success)$'

Expected Output: Step added with condition if_regex

Step 8: Execute if the previous output does NOT match a regex (here: fewer than 5 active services)

workflow-manager add my-workflow './alert-few-services.sh' --if-not-regex '(service.*active.*){5,}'

Expected Output: Step added with condition if_not_regex

Note

Quantifiers such as {5,} let you use a regex as a crude counter.

Step 9: Execute if a field in the previous step's JSON output equals a value

workflow-manager add my-workflow './handle-connected.sh' --if-json-path '$.status="connected"'

Expected Output: Step added with condition if_json_path_equals

Note

The previous output must be valid JSON. The CLI flag is --if-json-path; the serde tag written into the profile is if_json_path_equals.

Step 10: Execute if a nested JSON field, reached through an array index, matches

workflow-manager add my-workflow './finland-detected.sh' --if-json-path '$.data.records[0].country_name="Finland"'

Expected Output: Step added with condition if_json_path_equals

Note

Supported path syntax: $ root, .field, [index]. Recursive descent (..) and wildcards (*) are NOT supported.

Step 11: Execute if a JSON boolean is true

workflow-manager add my-workflow './ip-online.sh' --if-json-path '$.ip_connectivity=true'

Expected Output: Step added with condition if_json_path_equals

Note

Booleans and numbers are written unquoted: =true, =2. Strings need quotes: ="ok".

Step 12: Execute only when a live system state matches (independent of the previous step)

workflow-manager add my-workflow './tor-ready.sh' --if-state tor_running --if-state-expect true

Expected Output: Step added with condition if_state

Note

--if-state-expect takes JSON (true, false, 60, "wireguard"). Run `workflow-manager state` to see the state names.

Step 13: Execute based on a probe function defined in the profile

workflow-manager add my-workflow './skip-if-hardened.sh' --if-probe is_hardened --if-probe-expect false

Expected Output: Step added with condition if_probe

Note

The probe name must exist in the profile's `probes` block, see the Probe Functions category. --if-probe-expect takes JSON.

Step 14: Execute if a boolean expression over states and probes evaluates true

workflow-manager add my-workflow './low-score-fix.sh' --if-expression "state.authenticated == true && state.security_score < 60"

Expected Output: Step added with condition if_expression

Note

Grammar: atoms are probe('NAME') or state.NAME, operators == != < > <= >=, combined with && and || and negated with !. There is no contains() function and no step_id or workflow_status variable.

Scenario 5: Workflow Execution

Run workflows and control what happens when a step fails.

Step 1: List the steps that would run, without executing anything

workflow-manager run my-workflow --dry-run

Expected Output: Dry run: 5 steps would be executed

Note

Conditions that depend on previous output cannot be fully resolved in a dry run.

Step 2: Execute every step in order, evaluating each condition first

workflow-manager run my-workflow

Expected Output: Workflow completed: Success (5/5 steps, 2 skipped by condition)

Note

Steps run real shell commands. Read the template with `show` before running one you did not write.

Step 3: Keep going after a failed step instead of stopping

workflow-manager run my-workflow --kill-policy continue

Expected Output: Workflow completed with failures (step 3 failed, remaining steps executed)

Note

Values: stop (default and safest), continue, skip_remaining. This overrides global_settings.kill_policy for this run.

Step 4: Raise the per-run log cap for a very chatty workflow

workflow-manager run my-workflow --max-log-size 20971520

Expected Output: Workflow completed: Success

Note

Bytes, not megabytes. Default is 10485760 (10 MB). Overrides global_settings.max_log_size.

Step 5: Run unattended in text mode, skipping the [y/N] prompt on confirm:true steps

workflow-manager run my-workflow --yes

Expected Output: Workflow completed: Success (no confirmation prompts shown)

Note

--yes (-y) approves EVERY confirm:true step, including destructive ones, with no prompt. It only affects text mode: the [y/N] prompt is text mode only.

Step 6: Run non-interactively and get a machine readable result (this is how the dashboard runs workflows)

workflow-manager run my-workflow --json

Expected Output: JSON envelope with per-step status, exit codes and durations

Note

--json is non-interactive: confirm:true steps run WITHOUT any prompt and without --yes. Do not rely on confirm as a safety gate for automated runs.

Scenario 6: Managing Individual Steps

Update or delete a single step. Step IDs come from `show`.

Step 1: Find the step IDs before editing

workflow-manager show my-workflow

Expected Output: Step 1: echo Hello Step 2: ./backup.sh Step 3: ./cleanup.sh

Step 2: Replace step 1's command and timeout

workflow-manager update my-workflow 1 'echo Updated Command' -t 300

Expected Output: Step 1 updated successfully

Note

The command positional is optional: pass only the flags you want to change.

Step 3: Change a step's condition and its description in place

workflow-manager update my-workflow 2 --if-contains 'ERROR' --description 'Alert on error'

Expected Output: Step 2 updated successfully (condition if_contains)

Note

update accepts the same condition flags as add: -c, --if-contains, --if-not-contains, --if-equals, --if-regex, --if-not-regex, --if-json-path, --if-probe(-expect), --if-state(-expect), --if-expression.

Step 4: Drop a step's condition payload and reset it to a plain condition type

workflow-manager update my-workflow 2 --clear-condition -c always

Expected Output: Step 2 updated successfully (condition always)

Note

--clear-condition removes the stored pattern/path/probe value. Without it, switching from if_contains back to always can leave a stale payload.

Step 5: Update a pause step's message and require confirmation

workflow-manager update my-workflow 3 -m 'Verify the archive, then press Enter' --confirm true

Expected Output: Step 3 updated successfully

Note

-m/--message applies to pause steps, --profile applies to include steps, the command positional applies to command steps.

Step 6: Remove step 2 from the workflow

workflow-manager delete-step my-workflow 2

Expected Output: Step 2 deleted from 'my-workflow'

Note

Steps are renumbered after a delete, so re-run `show` before deleting another one.

Scenario 7: Pause and Include Steps

Two non-command step types: a pause that waits for the user, and an include that expands another profile inline.

Step 1: Add an interactive pause point

workflow-manager pause my-workflow -m 'Review results before continuing'

Expected Output: Pause step added to template 'my-workflow'

Note

The run blocks until the user presses Enter, so a paused workflow is not suitable for unattended or --json runs.

Step 2: Pause only when the previous step succeeded

workflow-manager pause backup -m 'Verify backup integrity' -c if_success

Expected Output: Pause step added with condition if_success

Step 3: Compose another profile into this template

workflow-manager include my-workflow shared-prereqs --description 'Reusable setup steps'

Expected Output: Include step 3 added to template 'my-workflow'

Note

The named profile must exist. Its steps are expanded in place when the workflow runs.

Step 4: Pull in a recovery profile only when the previous step failed

workflow-manager include my-workflow tor-recovery -c if_fail

Expected Output: Include step added with condition if_fail

Note

include accepts -c/--condition plus --if-probe(-expect), --if-state(-expect) and --if-expression.

Scenario 8: Efficient Batch Building

Build a complete workflow in one shell line by chaining create and add with &&.

Step 1: Create a template and add three conditional steps in one chain

workflow-manager create tor-recovery && workflow-manager add tor-recovery './routing-switch recover internet' -t 60 && workflow-manager add tor-recovery './health-control net-check' -c if_success && workflow-manager add tor-recovery './tor-switch start' -c if_success -t 120

Expected Output: Template created, 3 steps added successfully

Note

`create` must come first: `add` refuses to run against a template that does not exist.

Step 2: Build a 4 step diagnostics workflow

workflow-manager create diagnostics && workflow-manager add diagnostics './health-control net-check' -t 30 && workflow-manager add diagnostics './tor-switch tor-status' -c if_success && workflow-manager add diagnostics './dns-leak test' -c if_success && workflow-manager add diagnostics './integrity-check check-all' -c if_success

Expected Output: diagnostics created, 4 steps added

Step 3: Mix command steps and a pause to build an interactive backup workflow

workflow-manager create backup && workflow-manager add backup 'tar czf backup.tar.gz /data' && workflow-manager pause backup -m 'Check backup size' -c if_success && workflow-manager add backup 'rsync backup.tar.gz remote:/backups' -c if_success && workflow-manager add backup 'rm backup.tar.gz' -c if_success

Expected Output: Backup workflow with 4 steps created

Scenario 9: Real-World Kodachi Workflows

End to end workflows built from real Kodachi hook commands.

Step 1: Fetch the public IP and branch on the geolocated country

workflow-manager create ip-verify && workflow-manager add ip-verify 'ip-fetch --json' -t 60 && workflow-manager add ip-verify 'echo Finland detected' --if-json-path '$.data.records[0].country_name="Finland"' && workflow-manager run ip-verify

Expected Output: Finland detected

Note

Needs network access. ip-fetch does not need root.

Step 2: Confirm there is a valid authenticated session

workflow-manager create auth-check && workflow-manager add auth-check 'online-auth check-login --json' -t 30 && workflow-manager add auth-check 'echo Session valid' --if-contains 'valid' && workflow-manager run auth-check

Expected Output: Session valid

Note

Many hook commands refuse to run without an authenticated session, so this is a useful first step in a longer workflow.

Step 3: Chain a network check into a routing status check

workflow-manager create health-audit && workflow-manager add health-audit 'sudo health-control net-check --json' -t 60 && workflow-manager add health-audit 'echo Network online' --if-json-path '$.ip_connectivity=true' && workflow-manager add health-audit 'sudo routing-switch status --json' -c if_success -t 30 && workflow-manager run health-audit

Expected Output: Network online

Note

The steps themselves need root here (health-control and routing-switch do), workflow-manager does not.

Step 4: Verify the Tor daemon is up and answering

workflow-manager create tor-verify && workflow-manager add tor-verify 'sudo tor-switch get-tor-status --json' -t 30 && workflow-manager add tor-verify 'echo Tor responding' --if-json-path '$.data.is_responding=true' && workflow-manager run tor-verify

Expected Output: Tor responding

Note

Requires Tor to be installed and its control port reachable.

Scenario 10: Prerequisites

A profile can declare system state that must hold before any step runs. The block is `prerequisites` and it has exactly three fields: `required` (array), `optional` (array, warnings only) and `on_failure`. There is no flat form: keys such as "authenticated": true at the top of the block do not deserialize.

Step 1: Validate a profile's prerequisites without running the workflow

workflow-manager prereq check initial_terminal_setup_wireguard_torrify

Expected Output: All prerequisites met, or the list of failed checks with their error messages

Note

`prereq validate <TEMPLATE>` is accepted as an alias for `prereq check`.

Step 2: Read the exact prerequisites shape the runner expects

workflow-manager show initial_terminal_setup_wireguard_torrify --json

Expected Output: The profile JSON, including the prerequisites block

Note

The shape is: "prerequisites": {"required": [{"check": "state.online", "expect": true, "error": "Internet connection required"}, {"check": "state.torrify", "expect": false, "error": "System must not be torrified"}], "on_failure": "abort"}. Each entry needs `check` and `error`; `expect` is optional and defaults to a truthy check.

Step 3: List the state names usable as a prerequisite `check` value

workflow-manager state

Expected Output: A JSON object of all system states, for example {"online": true, "torrify": false, "tor_running": true, ...}

Note

A check is written as "state.<NAME>" (the bare "<NAME>" form is also accepted). Optional checks go in the `optional` array: they warn but never block.

Step 4: See how on_failure behaves before committing to a real run

workflow-manager run initial_terminal_setup_wireguard_torrify --dry-run

Expected Output: Prerequisites validated, or 'Prerequisites not met' followed by the failed checks

Note

on_failure accepts "abort" (default) and "warn". Only "abort" stops the run: any other value, including a typo, lets the workflow continue. If the profile has no prerequisites block, no checks are performed.

Scenario 11: System State Checking

Query the live system state that conditions, probes and prerequisites read from.

Step 1: Show every system state as JSON

workflow-manager state

Expected Output: {"online": true, "torrify": false, "dnscrypt": true, "routing_mode": "wireguard", "tor_running": true}

Step 2: Query a single state field (cheaper: only that field is probed)

workflow-manager state online

Expected Output: {"state": "online", "value": true}

Note

20 states are available: online, routing_mode, vpn_connected, dnscrypt, ipv6_disabled, dns_kodachi_managed, firewall_active, kill_switch_armed, network_hardened, disk_encrypted, security_score, torrify, tor_running, tor_dns_active, tor_verified, mac_spoofing, bluetooth_enabled, wifi_enabled, authenticated, tor_instances_count.

Step 3: Read a numeric state, for use with --if-expression comparisons

workflow-manager state security_score --json

Expected Output: {"state": "security_score", "value": 62}

Note

Numeric states support < > <= >= inside --if-expression, for example "state.security_score < 60".

Scenario 12: Probe Functions (profile JSON)

A probe is a named, reusable check declared in the profile's `probes` block and referenced by --if-probe. A ProbeDefinition has exactly three fields: `probe_type` ("builtin", "expression" or "count"), `expression` (required for all three types) and an optional `description`. There is no `check` field.

Step 1: Reference a probe from a step condition

workflow-manager add my-workflow 'echo Already hardened, skipping' --if-probe is_hardened --if-probe-expect true

Expected Output: Step added with condition if_probe

Note

Add the probes block to config/profiles/my-workflow.json first: "probes": {"harden_count": {"probe_type": "count", "expression": "count('harden', previous_output)", "description": "Count 'harden' occurrences"}, "is_hardened": {"probe_type": "expression", "expression": "probe('harden_count') >= 4"}}

Step 2: Use the one builtin probe: a live Tor connection check

workflow-manager add my-workflow './use-tor-path.sh' --if-probe tor_up --if-probe-expect true

Expected Output: Step added with condition if_probe

Note

Declare it as "tor_up": {"probe_type": "builtin", "expression": "tor_connection_check"}. tor_connection_check is the ONLY builtin: any other expression under probe_type "builtin" fails at run time with 'Unknown builtin probe'. It shells out to ip-fetch check-tor, so it needs network access, and its result is cached for the run.

Step 3: Confirm the probes block deserializes the way the runner expects

workflow-manager show my-workflow --json

Expected Output: The profile JSON with its `probes` map, each entry showing probe_type and expression

Note

probe_type "count" evaluates count('<TEXT>', previous_output) and returns a number. probe_type "expression" evaluates the same grammar as --if-expression, so it can reference other probes with probe('NAME') and states with state.NAME.

Scenario 13: Global Settings (profile JSON)

The `global_settings` block sets workflow wide defaults: kill_policy ("stop" by default), continue_policy (false by default), max_log_size (10485760 bytes), default_timeout (seconds) and working_dir. Step level values override them.

Step 1: Read the global_settings block of an existing profile

workflow-manager show policy-demo --json

Expected Output: "global_settings": {"kill_policy": "stop", "continue_policy": false, "max_log_size": 10485760, "default_timeout": 300, "working_dir": "."}

Note

continue_policy is the master override: when true the workflow always continues and kill_policy is ignored. When false, kill_policy decides: "stop" ends the run, "continue" carries on, "skip_remaining" skips the rest without failing.

Step 2: Force the safest failure policy for one run, whatever the profile says

workflow-manager run policy-demo --kill-policy stop

Expected Output: Workflow stopped at the first failing step

Note

For security critical workflows use kill_policy "stop" with continue_policy false. For recovery workflows, "continue" is usually what you want.

Step 3: Give one step its own timeout, overriding default_timeout

workflow-manager add slow-scan './deep-scan.sh' -t 60

Expected Output: Step added with timeout 60s

Note

A step with no timeout inherits global_settings.default_timeout. A timed out command is sent SIGTERM, then SIGKILL if it does not exit.

Scenario 14: Condition Reference (profile JSON)

The 12 condition tags the runner accepts, exactly as they are written in the profile's `condition` object. Anything else fails to deserialize: always, if_success, if_fail, if_contains, if_not_contains, if_equals, if_regex, if_not_regex, if_json_path_equals, if_probe, if_state, if_expression.

Step 1: See the condition object the CLI flags actually write

workflow-manager show my-workflow --json

Expected Output: "condition": {"type": "if_contains", "pattern": "SUCCESS"}

Note

Field per tag: always / if_success / if_fail take none. if_contains, if_not_contains, if_regex, if_not_regex take `pattern`. if_equals takes `value`. if_json_path_equals takes `path` and `value`. if_probe takes `probe` and `expect`. if_state takes `state` and `expect`. if_expression takes `expression`.

Step 2: Substring matching, written as if_contains / if_not_contains

workflow-manager add my-workflow './alert.sh' --if-contains 'SUCCESS'

Expected Output: Step added: {"type": "if_contains", "pattern": "SUCCESS"}

Note

This is a plain substring test, case sensitive. It is NOT glob: * and ? are literal characters here. For wildcards use if_regex.

Step 3: JSON path equality, written as if_json_path_equals

workflow-manager add my-workflow './connected.sh' --if-json-path '$.status.connected=true'

Expected Output: Step added: {"type": "if_json_path_equals", "path": "$.status.connected", "value": true}

Note

The stored field is `value`, not `expect`. Path syntax is a strict subset: $ root, .field, [index]. Recursive descent (..) and wildcards are not supported.

Step 4: Boolean expression, written as if_expression

workflow-manager add my-workflow './score-fix.sh' --if-expression "probe('harden_count') >= 4 || state.security_score < 60"

Expected Output: Step added: {"type": "if_expression", "expression": "probe('harden_count') >= 4 || state.security_score < 60"}

Note

Atoms: probe('NAME') and state.NAME. Operators: == != < > <= >=, joined with && and ||, negated with a leading !. Comparisons on booleans and strings support only == and !=.

Scenario 15: Step Reference (profile JSON)

A step object has two required fields, `id` and `type` ("command", "pause" or "include"), plus `condition` (defaults to always if absent). Everything else is optional: cmd, description, timeout, confirm, env_allowlist, working_dir, nice_level, message, profile, parameters, and the ColonyOps hints cell_store, score_impact and colonyops_safety.

Step 1: Read a complete step object as the runner stores it

workflow-manager show my-workflow --json

Expected Output: {"id": 1, "type": "command", "cmd": "health-control net-check", "description": "Check network connectivity", "condition": {"type": "always"}, "timeout": 60, "confirm": false, "nice_level": 0}

Note

`description` is optional, not required. A command step needs `cmd`, a pause step needs `message`, an include step needs `profile` (and may carry `parameters`).

Step 2: Write the confirm and nice_level step fields from the CLI

workflow-manager add my-workflow './panic.sh' --confirm --nice-level -10

Expected Output: Step added: {"confirm": true, "nice_level": -10}

Note

confirm:true prompts [y/N] in text mode only, and `run --yes` skips that prompt. A --json run never prompts. nice_level goes from -20 to 19; negative values need root.

Step 3: Write the working_dir and env_allowlist step fields

workflow-manager add my-workflow './report.sh' -w /opt/kodachi/dashboard/hooks --env 'HOME,PATH'

Expected Output: Step added: {"working_dir": "/opt/kodachi/dashboard/hooks", "env_allowlist": ["HOME", "PATH"]}

Note

A step without working_dir inherits global_settings.working_dir. env_allowlist is a whitelist: variables not listed are not passed to the step.

Step 4: Write an include step, which expands another profile at run time

workflow-manager include my-workflow base-recovery-sequence

Expected Output: Step added: {"type": "include", "profile": "base-recovery-sequence", "condition": {"type": "always"}}

Note

Include lets you keep a library of small reusable profiles and compose them. The included profile must already exist.

Environment Variables

Variable	Description	Default	Values
`NO_COLOR`	Disable all colored output when set	unset	1\|true\|yes (any value disables color)

Exit Codes

Code	Description
0	Success
1	General error
2	Invalid arguments
3	Permission denied
4	Network error
5	File not found