ulu exec

Execute agents, commands, workflows, and pipelines. Run validation directly from the terminal.

Alias: ulu x

7 commands

Parent Options

These options apply to all subcommands in this group.

--local-definitions <dir>

Local YAML definitions directory

--project <name>

Project name for result tracking

--no-tracking

Disable validation service submission

--no-safety-warnings

Suppress pre-run risk warnings, incomplete-analysis advisories, and sensitive-path runtime advisories

--hash <sha256:...>

Pin the expected definition YAML hash — verified fail-closed before execution; exit code 4 on mismatch. Accepted by every exec subcommand

--prompt-hash <sha256:...>

Pin the expected rendered-prompt hash (run/agent/command only — workflows and pipelines have no rendered prompt)

exec run

Execute a definition (auto-detects type from registry).

Arguments

namerequired

Definition name

targetrequired

Target directory to analyze

Options

-m, --model <model>

Model override (alias, tier, or provider:modelId)

-p, --prompt <text>

Operator directive or context

ulu exec run <name> <target>

exec agent

Execute one or more agent definitions. Runs in parallel when multiple agents are specified.

Arguments

namesrequired

Agent name(s) — space-separated for parallel execution

Options

-t, --target <path>

Target directory to analyze

-m, --model <model>

Model override (alias, tier, or provider:modelId)

-p, --prompt <text>

Operator directive or context

--max-tokens <n>

Maximum response tokens

--max-steps <n>

Maximum tool loop iterations

Default: 50

--temperature <n>

Generation temperature 0-1

Default: 0

--timeout <ms>

Execution timeout in milliseconds

--threshold-pass <n>

Pass threshold score

--threshold-warn <n>

Warning threshold score

--report <path>

Write raw agent output report to file (single agent only)

--features-list <path>

Write structured features/recommendations to file (single agent only)

ulu exec agent <names...>

Examples

# Run code-validator with a specific model
ulu exec agent code-validator -t ./src --model sonnet --project my-project

# Run with operator context
ulu exec agent security-analyst -t ./src -p "Focus on JWT handling"

# Run two agents in parallel
ulu exec agent code-validator test-architect -t ./src

exec command

Execute a saved command configuration.

Arguments

namerequired

Command definition name

targetrequired

Target directory

Options

-m, --model <model>

Model override (overrides command definition default)

-p, --prompt <text>

Operator directive or context

ulu exec command <name> <target>

exec workflow

Execute a multi-phase workflow.

Arguments

namerequired

Workflow definition name

targetrequired

Target directory

Options

-m, --model <model>

Model override for all phases

-p, --prompt <text>

Operator directive or context

ulu exec workflow <name> <target>

Examples

# Run post-implementation workflow
ulu exec workflow post-implementation ./ --project my-project

# Run without tracking results
ulu exec workflow ship ./packages/api --no-tracking

exec pipeline

Execute a multi-stage pipeline.

Arguments

namerequired

Pipeline definition name

targetrequired

Target directory

Options

-m, --model <model>

Model override for all stages

-p, --prompt <text>

Operator directive or context

ulu exec pipeline <name> <target>

Examples

# Run foundations pipeline
ulu exec pipeline foundations ./ --project my-project

# Run without tracking results
ulu exec pipeline ship ./packages/api --no-tracking

exec list

List available definitions from registry and local sources.

Options

-t, --type <type>

Filter by type (agent, command, workflow, pipeline)

-d, --domain <domain>

Filter by domain

ulu exec list

exec describe

Inspect a definition's metadata and interface.

Arguments

namerequired

Definition name

ulu exec describe <name>

Examples

# View agent metadata and interface
ulu exec describe code-validator