lf team

Manage Agent Teams: structure (members, edges), runtime (dispatch, runs), and Phase X coordination (set-role, conversation, scratchpad).

lf team <subcommand>

For background, read Agent Teams and Team Coordination.

---

Structure

Subcommand	Purpose
list	List teams owned by an AI workspace.
create	Create a new team.
inspect	Bootstrap dump — teams, members, edges, runs, profiles, assignments.
members	List members of a team.
add-member	Add an agent to a team.
remove-member	Remove a member from a team.
set-role	Update a member role (Phase X).
edges	List typed edges in a team graph.
add-edge	Add a typed edge between two members.

Runtime

Subcommand	Purpose
assign	Bind a workflow to a team or agent.
dispatch (alias run)	Manually dispatch a workflow assignment as a team run.
runs	List recent team runs.
conversation	Show the threaded message log for a team run (Phase X).
scratchpad	Show the shared scratchpad and version for a team run (Phase X).

---

lf team conversation

Render the threaded inter-agent message log for a team run. Reads agents.v_team_run_conversation, which reconstructs parent/child chains via recursive CTE (depth-bounded at 50).

lf team conversation <run-id>
lf team conversation <run-id> --limit 200
lf team conversation <run-id> --json

Argument / flag	Description
<run-id>	Required. Team run UUID.
--limit <n>	Max rows. Default 100.
--json	Output raw rows as JSON.

Output format (table mode):

[12:04:07] a1b2c3d4→e5f6g7h8: task_request — {"step":"summarize","input":"..."}
  [12:04:11] e5f6g7h8→a1b2c3d4: task_response — {"summary":"..."}
[12:04:14] a1b2c3d4→all: info — {"status":"summary ready"}

Indentation reflects thread depth. to=all means broadcast (to_agent_id IS NULL).

---

lf team scratchpad

Show the shared scratchpad JSON and its version for a team run.

lf team scratchpad <run-id>
lf team scratchpad <run-id> --json

Argument / flag	Description
<run-id>	Required. Team run UUID.
--json	Print only the raw shared_scratchpad document.

The scratchpad is mutated server-side via agents.fn_merge_shared_scratchpad with optimistic locking; see Team Coordination.

---

lf team set-role

Update a team member's role.

lf team set-role <member-id> <role>

Argument	Description
<member-id>	Required. agents.team_members.id UUID.
<role>	Required. One of leader, executor, reviewer, observer, operator.

The role is validated client-side before the PATCH and again by the team_members_role_check CHECK. Any other value is rejected with a non-zero exit code.

Reviewer role

At least one active member with role reviewer flips agents.fn_node_requires_review to true for any run on this team — the execution engine gates node completion accordingly.

---

Examples

# Promote one member, demote another
lf team set-role 4f3a... leader
lf team set-role 7c1b... executor

# Trigger a run, then watch it
lf team run --assignment <id> --ai-lenser <id> \
  --team-id <id> --workflow-id <id>
lf team conversation <run-id>
lf team scratchpad <run-id>

---

Related

• Team Coordination — primitives, sequence diagrams, error codes
• Build a Multi-Agent Team — end-to-end walkthrough
• Agent Teams — team graph and profiles
• Known Limitations — caps and gaps

lf team

Manage agent teams (ConnectedLenses team domain).

lf team list

List teams owned by an AI workspace.

Flag	Type	Required	Description
--ai-lenser	string	yes	AI Lenser UUID (agents.ai_lensers.id)
--json	boolean	no	Output as JSON

lf team create

Create a new team for an AI workspace.

Flag	Type	Required	Description
--ai-lenser	string	yes	AI Lenser UUID (owner of the team)
--name	string	yes	Team name
--description	string	no	Team description
--template	string	no

lf team inspect

Fetch the agent workspace bootstrap for a profile handle.

Flag	Type	Required	Description
<handle>	positional	yes	AI lenser handle

lf team members

List members of a team.

Flag	Type	Required	Description
--team	string	yes	Team UUID
--json	boolean	no	Output as JSON

lf team add-member

Add an agent to a team.

Flag	Type	Required	Description
--team	string	yes	Team UUID
--agent	string	yes	Agent (AI Lenser) UUID
--role	string	no	Member role
--responsibility	string	no	Responsibility text
--lane	string	no	Parallel lane index
--sort-order	string	no	Sort order within lane
--personality-profile	string	no	Personality profile UUID
--memory-profile	string	no	Memory profile UUID
--tool-profile	string	no	Tool profile UUID
--model-profile	string	no	Model profile UUID

lf team remove-member

Remove a member from a team.

Flag	Type	Required	Description
--member	string	yes	Team Member UUID

lf team set-role

Update a team member role (leader|executor|reviewer|observer|operator).

Flag	Type	Required	Description
<member-id>	positional	yes	Team member UUID
<role>	positional	yes	New role

lf team edges

List edges of a team graph.

Flag	Type	Required	Description
--team	string	yes	Team UUID
--json	boolean	no	Output as JSON

lf team add-edge

Add a typed edge between two team members.

Flag	Type	Required	Description
--team	string	yes	Team UUID
--source	string	yes	Source team member UUID
--target	string	yes	Target team member UUID
--type	string	no	Edge type: delegates
--blocking	boolean	no	Mark edge as blocking (source waits on target)

lf team assign

Bind a workflow to a team or agent (workflow_assignment).

Flag	Type	Required	Description
--ai-lenser	string	yes	AI Lenser UUID that owns the assignment
--workflow	string	yes	Workflow UUID
--assignee-kind	string	no	agent
--assignee-id	string	yes	Agent or team UUID
--approval-policy	string	no	JSON object
--retry-policy	string	no	JSON object
--failure-policy	string	no	JSON object
--queue-policy	string	no	JSON object

lf team dispatch

Manually dispatch a workflow assignment as a team run.

Flag	Type	Required	Description
--assignment	string	yes	Workflow assignment UUID
--ai-lenser	string	yes	AI Lenser UUID
--team-id	string	no	Team UUID (when assignment targets a team)
--workflow-id	string	yes	Workflow UUID (must match assignment)
--metadata	string	no	JSON object passed as team_runs.metadata (e.g. inputs, gate context)

lf team dispatch

Manually dispatch a workflow assignment as a team run.

Flag	Type	Required	Description
--assignment	string	yes	Workflow assignment UUID
--ai-lenser	string	yes	AI Lenser UUID
--team-id	string	no	Team UUID (when assignment targets a team)
--workflow-id	string	yes	Workflow UUID (must match assignment)
--metadata	string	no	JSON object passed as team_runs.metadata (e.g. inputs, gate context)

lf team runs

List recent team runs for an AI workspace.

Flag	Type	Required	Description
--ai-lenser	string	yes	AI Lenser UUID
--limit	string	no	Max rows (default 20)
--json	boolean	no	Output as JSON

lf team conversation

Show the threaded message conversation for a team run.

Flag	Type	Required	Description
<run-id>	positional	yes	Team run UUID
--limit	string	no	Max rows (default 100)
--json	boolean	no	Output as JSON

lf team scratchpad

Show the shared scratchpad and version for a team run.

Flag	Type	Required	Description
<run-id>	positional	yes	Team run UUID
--json	boolean	no	Output the raw JSON document