agentleFS
Sign inStart free
VodouAIVodouAI/OS/skills/templates/ship-a-feature/SKILL.md

ship-a-feature

Skill1 starsChanged 4 months agoApache-2.0View on GitHub

workflow template — PM specs, engineer implements, reviewer approves, deployer ships; pre-installed with the board, activates in Phase 2 when the workflow_template kind enum lands

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

Create a free account to mount
Files
SKILL.md

ship-a-feature

A 4-stage pipeline for shipping a feature end-to-end: spec → implement → review → deploy. The dispatcher auto-advances tasks through these stages when each stage's required_artifacts validate.

Phase 1 status: pre-installed but inactive. The kernel migration's tasks.workflow_template_id + current_step_key columns hold the binding; Phase 2 lands the loader that reads this file's AGENT_ACTIONS_N blocks into board.db::board_templates.

Stages

Stage 1 — Specify

<!-- AGENT_ACTIONS_1: { "step_key": "spec", "assignee_default": "pm", "skills": ["board-worker", "product-spec"], "workspace": "scratch", "required_artifacts": { "metadata.spec_doc_path": "\\.md$", "metadata.acceptance_criteria_count": "[1-9][0-9]*" }, "on_success": "implement", "on_failure": "block", "max_runtime_seconds": 1800, "budget_usd_cap": 1.50 } -->

The PM (or PM-equivalent subagent) writes the spec: goal, approach, acceptance criteria, success metrics. Output is a markdown file. The dispatcher checks the file path matches .md$ and the acceptance_criteria_count is ≥1 before advancing.

Stage 2 — Implement

<!-- AGENT_ACTIONS_2: { "step_key": "implement", "assignee_default": "engineer", "skills": ["board-worker", "engineer-typescript", "engineer-rust"], "workspace": "worktree", "required_artifacts": { "metadata.pr_url": "^https://github\\.com/", "metadata.tests_run": "[1-9][0-9]*", "metadata.changed_files_count": "[1-9][0-9]*" }, "on_success": "review", "on_failure": "implement", "max_runtime_seconds": 7200, "max_retries": 2, "budget_usd_cap": 8.00 } -->

The engineer (or engineer-equivalent subagent) writes the code in a git worktree, runs tests, opens a PR. Required artifacts: metadata.pr_url matching GitHub, tests_run ≥ 1, changed_files_count ≥ 1. On failure, retries once (back to implement); on second failure, falls to block for human review.

Stage 3 — Review

<!-- AGENT_ACTIONS_3: { "step_key": "review", "assignee_default": "reviewer", "skills": ["board-worker", "security-pr-audit", "github-code-review"], "workspace": "scratch", "required_artifacts": { "metadata.approved": "true", "metadata.security_audit_passed": "true" }, "on_success": "deploy", "on_failure": "implement", "max_runtime_seconds": 3600, "budget_usd_cap": 2.00, "requires_approval_on": ["running→done"] } -->

The reviewer (or reviewer-equivalent subagent) does the editorial + security pass. Must explicitly set metadata.approved=true AND metadata.security_audit_passed=true. Approval gate fires before done so a human signs off. On rejection, kicks back to implement (engineer iterates).

Stage 4 — Deploy

<!-- AGENT_ACTIONS_4: { "step_key": "deploy", "assignee_default": "deployer", "skills": ["board-worker", "deploy-prod"], "workspace": "scratch", "required_artifacts": { "metadata.deploy_status": "^success$", "metadata.deploy_url": "^https?://" }, "on_success": null, "on_failure": "block", "max_runtime_seconds": 1800, "budget_usd_cap": 1.00 } -->

The deployer ships it. Requires deploy_status=success AND a deploy_url. On failure, blocks for human investigation.

Pipeline total budget cap (recommendation)

Sum of stage caps: ~$12.50 per feature run. The board enforces per-stage caps; cumulative spend across stages rolls up via SUM(task_runs.usd_spent) for the parent task.

When to use this template vs not

✓ Use it when the work follows the standard ship-a-feature shape: discrete spec, code change, review, deploy.

✗ Skip it when:

  • It's research-only (no code change → use daily-research-brief instead)
  • It's a documentation-only change (no review needed → just create a task with assignee=writer)
  • It's a hotfix (the pipeline is too slow for incidents → create direct + auto-approve)

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free