- Home
- Skills
- Productivity
- Project Kit
Project Kit
Generate a CLAUDE.md from what your repo really contains, keep it current, and decide when a task is worth a subagent — without burning quota on delegation that doesn't pay.
$5
Works with the AI Agents you already use
Project Kit
Example session with this skill installed
I just cloned this repo and it has no instructions. Write a CLAUDE.md for it and tell me if I should use a subagent to map the data flow.
- Read your context and instructions
- Compiled the project kit
I have analyzed the repository structure. It appears to be a TypeScript project using Vite and Vitest. I have drafted a CLAUDE.md based on the discovered scripts and patterns. Regarding your request to map data flow: this is a high-context architectural task involving multiple modules, so I recommend delegating this to a subagent using a long-context model.
CLAUDE.md Draft
# Project Guide: Vite-TS-App
## Build & Test Commands
- Build: `npm run build`
- Dev: `npm run dev`
- Test: `npm run test`
- Lint: `npm run lint`
## Project Structure
- `src/components`: UI components (React)
- `src/hooks`: Custom React hooks
- `src/services`: API logic and data fetching
- `src/types`: TypeScript interfaces/types
## Style & Patterns
- Use functional components with hooks.
- Prefer `interface` over `type` for public APIs.
- All new features require a corresponding `.test.ts` file.
Delegation Recommendation
Decision: Delegate to Subagent
Recommended Model: Claude 3.5 Sonnet (or equivalent long-context model)
Briefing Template:
Task: Map data flow from API services to UI components.
Focus: TraceUserService.tsthroughuseAuthhook toNavbarcomponent.
Output: Return a concise summary of the lifecycle, including file paths and key function names. Do not provide raw logs.
Next steps
- Approve the
CLAUDE.mdcontent to save it to the root directory. - Initialize the subagent using the provided briefing template.
- Run
npm run testto verify the current environment state.
Connects securely to your tools. The creator never sees your data.
About this skill
Onboarding a new agent to a legacy repository often leads to hallucinated build commands and ignored style guides. This skill anchors agent behavior to the reality of your codebase by generating and maintaining precise instruction files.
What it does
- Repo Analysis scans your project structure to identify actual build, test, and lint commands.
- CLAUDE.md Generation creates a project-specific guide including directory maps, tech stacks, and "Definition of Done."
- Instruction Auditing reviews existing guide files against current repository state to fix drift or missing rules.
- Delegation Logic evaluates task complexity to decide if a job stays inline or moves to a subagent.
- Subagent Briefing generates structured prompts that force subagents to return concise summaries instead of unparsed logs.
How it works
- Trigger the skill using
initordelegatecommands when starting a repo or facing a large task. - The agent crawls the file system to extract real paths, scripts, and configuration patterns.
- For instruction updates, the skill proposes a merge rather than overwriting existing files.
- For heavy tasks, it outputs a model recommendation and a specific briefing template.
Frameworks & tools
This skill is optimized for environments using CLAUDE.md or AGENTS.md conventions. It works across any stack by reading standard package managers, test runners, and configuration files like package.json, pyproject.toml, or Makefile.
Why this beats prompting it yourself
Generic prompts often result in vague advice like "write clean code." This skill forces the agent to use the exact commands found in your repo, preventing broken builds and ensuring subagents stay within scope without burning unnecessary tokens.
Use cases
- Generate a CLAUDE.md for an unfamiliar repository to start working immediately.
- Update stale project instructions after a major refactor or stack migration.
- Determine if a complex bug investigation justifies the cost of a subagent.
- Create high-density briefs for subagents to ensure actionable summaries.
Known limitations
Does not support overwriting existing files without manual approval. Effectiveness depends on the presence of standard configuration files for command detection.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Productivity.
Requires existing project files to detect commands. Will not overwrite instruction files automatically. Scope is limited to repo-based instructions and delegation logic.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 1 day ago