- Home
- Skills
- Bgauryy
- Octocode Mcp
- Octocode Implement
octocode-implement_skill
- TypeScript
725
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 bgauryy/octocode-mcp --skill octocode-implement- SKILL.md10.0 KB
Overview
This skill is an Implementation Agent that turns specification documents into safe, tested TypeScript changes. It emphasizes deep research before coding, matching existing patterns, and small incremental commits with explicit checkpoints. The workflow enforces spec validation, context mapping, research, implementation, and strict validation gates.
How this skill works
Given a spec document and project context, the agent validates the spec, maps the codebase using local search and LSP tools, and drafts a step-by-step plan. It performs targeted research to find existing patterns, creates tasks for each implementation step, applies minimal changes with tests, and produces a changes log in the .octocode/implement session folder.
When to use it
- When you have a written spec (.md) and need feature implementation in TypeScript
- When changes touch multiple files or require API/type updates
- When you must preserve existing architectural patterns and coding style
- When tests and compile-time safety are required before merge
- When you want explicit user checkpoints for ambiguous or high-impact decisions
Best practices
- Validate the spec fully before coding; STOP and ask if unclear
- Map the codebase (localViewStructure) and find similar implementations first
- Prefer types-first and small, incremental changes with tests
- Use TaskCreate/TaskUpdate to track and parallelize only independent work
- Run TypeScript compile, linter, and tests before marking tasks done
Example use cases
- Add a new API endpoint following existing service patterns with tests and docs
- Implement auth feature across backend and frontend with parallel tasks and integration checks
- Fix a complex bug that requires tracing call hierarchies and finding usages
- Extend an existing data model: update types, propagate changes, and add migration tests
- Prototype feature using external reference implementations from GitHub then adapt to local patterns
FAQ
Search using semantic variants and external references; if still unclear, stop and ask the user for guidance before inventing a new pattern.
When should I spawn parallel subagents?
Only when tasks are truly independent (different packages/modules or non-overlapping files). Use TaskCreate to define boundaries and merge changes sequentially with validation.