Stats
Actions
Tags
From dev-on-leash
Use to execute a single task from a plan (docs/plans/*.md). Reads the task's task-meta block, runs its verify command, and ticks the first checkbox of the task on success.
How this skill is triggered — by the user, by Claude, or both
Slash command
/dev-on-leash:execute-plan-taskThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
The user references a specific task in an existing plan and wants it executed mechanically — for example: "run T05 from the execution-harness plan". Use only when the plan + task id are known. Do NOT use to plan or design new work.
The user references a specific task in an existing plan and wants it executed mechanically — for example: "run T05 from the execution-harness plan". Use only when the plan + task id are known. Do NOT use to plan or design new work.
Resolve the plan file under docs/plans/<file>.md.
Invoke the harness runner:
python scripts/harness/run_task.py docs/plans/<file>.md <task_id>
The runner executes the task's verify: command and exits 0 if green. On exit 0, the first - [ ] checkbox of the task is replaced with - [x] in the plan file.
If verify fails (exit 1), DO NOT tick the checkbox by hand. Diagnose the failure, fix the underlying issue, re-run the skill.
If exit 2 (task id not found / usage error), the plan file or id is wrong — re-check the input.
--force or any flag to skip verify. The whole point is verify cannot be bypassed.Guides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.
npx claudepluginhub diogeneshfg/dev-on-leash --plugin dev-on-leash