- Home
- Skills
- Notedit
- Happy Skills
- Feature Pipeline
feature-pipeline_skill
- Python
278
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill notedit/happy-skills --skill feature-pipeline- SKILL.md3.0 KB
Overview
This skill executes implementation tasks directly from design documents that use markdown checkboxes. It reads the next task, performs the implementation loop, verifies acceptance criteria, and updates the task status in-place so no separate session files are required. It supports unattended runs, resuming interrupted work, and batch execution of queued tasks.
How this skill works
The skill scans a design markdown file for tasks formatted as checkbox items with metadata (priority, phase, files, deps, and subcriteria). It exposes commands to fetch the next task, mark tasks done or failed with reasons, and show overall status. In an execution loop it implements the listed files and criteria, updates the checklist when criteria pass, and records success or failure back into the same design document.
When to use it
- Implement features generated by a design analyzer or feature-analyzer output.
- Resume an interrupted implementation run using the same design markdown file.
- Batch execute a list of implementation tasks without interactive prompts.
- Automate step-by-step execution when tasks include file paths and acceptance criteria.
- Run unattended implementations that must make autonomous decisions following code patterns.
Best practices
- Keep tasks well-scoped with clear acceptance criteria and file paths to minimize ambiguity.
- Declare dependencies between tasks in the task metadata to preserve ordering.
- Use priority and phase tags to control execution order and grouping.
- Verify tests or criteria locally before marking a task done; use fail with a clear reason when blocked.
- Run status periodically to inspect progress and to resume from the first incomplete task.
Example use cases
- Implement a user model, associated tests, and mark subtasks as completed when validation and hashing are implemented.
- Create authentication utilities and API endpoints by following a chained task list with deps linking model and util tasks.
- Resume a long-running rollout after an interruption; the skill continues from the first unchecked task.
- Batch-execute small refactors across multiple files identified in a design document.
- Run unattended test-driven implementations where the agent applies patterns from the existing codebase.
FAQ
Use the provided commands to mark tasks done or fail; include a reason when failing so the design document records the blocker.
Can the skill run without human confirmation?
Yes. In unattended mode it makes autonomous decisions based on repository patterns and will mark tasks failed if blocked instead of stopping for questions.