Taskctl

Overview

Use taskctl from PATH.

Required toolchain for this skill:

• •taskctl (required)
• •gh (optional): Issue-driven state updates/comments
• •bddc (optional): behavior verification before completed

Core constraints:

• •status flow: pending -> in_progress -> completed
• •deleted removes task permanently
• •no self-dependency
• •no missing dependency target
• •no dependency cycle
• •blocked tasks cannot move to in_progress or completed

Workflow

1. •Choose a store file path.
2. •Create tasks with complete fields.
3. •Update owner, metadata, and dependencies.
4. •Run ready to identify executable tasks.
5. •Run dag or dag-ascii for graph review.
6. •Run validate before accepting task plan changes.
7. •If issue-driven, sync concise status to Issue via gh issue comment.

Command Patterns

# help
taskctl --help

# create task
taskctl --store <STORE_JSON> create \
  --subject "Run tests" \
  --description "Execute backend tests and collect logs" \
  --active-form "Running tests" \
  --metadata '{"priority":"P1","module":"quality"}'

# update task content, owner, dependencies, status
taskctl --store <STORE_JSON> update \
  --task-id <TASK_ID> \
  --owner qa@team \
  --metadata '{"priority":"P0","obsolete":null}' \
  --add-blocked-by <DEP_1>,<DEP_2> \
  --status in-progress

# inspect graph
taskctl --store <STORE_JSON> ready
taskctl --store <STORE_JSON> dag
taskctl --store <STORE_JSON> dag-ascii

# validate graph invariants
taskctl --store <STORE_JSON> validate

Field Semantics

• •subject: imperative task title.
• •description: context + acceptance criteria.
• •active_form: progressive action text.
• •owner: responsible human or agent.
• •metadata: object payload; null value deletes key on update.
• •add_blocked_by: dependency task IDs required before start.
• •add_blocks: task IDs blocked by current task.

Acceptance Checklist

1. •Ensure validate returns { "ok": true }.
2. •Ensure dag reflects intended module sequencing.
3. •Ensure ready includes only tasks with resolved dependencies.
4. •Include both JSON DAG and ASCII DAG when a human review is requested.