- Home
- Skills
- Technical Documentation
- README From Reality
Works with the AI Agents you already use
README From Reality
Generate a polished README from what the repository actually contains, with verified commands and marked unknowns.
Free
README From Reality
Example session with this skill installed
Generate a README from this repository evidence. Do not invent missing setup details.
package.json
- scripts:
dev: vite,build: vite build,test: vitest - dependencies: React, Vite, Zod
.env.example
VITE_API_URL=
Files
src/main.tsxsrc/api/client.tstests/vite.config.ts
The existing README only contains the project title.
- Read your context and instructions
- Compiled the readme from reality
README DRAFT
Project Name
A React application built with Vite. The repository includes a typed API client and Vitest-based tests.
Requirements
- Node.js compatible with the repository's current toolchain
- The repository's package manager/lockfile should be used for installation
Configuration
Create your local environment file from .env.example and provide:
VITE_API_URL=
Development
Use the repository's package manager to install dependencies, then run the dev script.
Available scripts
dev— starts Vite development modebuild— creates a production Vite buildtest— runs Vitest
DOCUMENTATION EVIDENCE
Verified
- React, Vite, Vitest, and Zod appear in
package.json. dev,build, andtestscripts exist.VITE_API_URLappears in.env.example.
Inferred
src/api/client.tssuggests the application communicates with an external API.
Missing / needs owner input
- Exact Node.js version.
- Product/business description.
- Deployment method.
Commands to verify before publishing
- Run install using the detected lockfile's package manager.
- Run
dev,build, andtest.
Update triggers
- Package scripts change.
- Environment schema changes.
- Deployment documentation is added.
Connects securely to your tools. The creator never sees your data.
About this skill
A README should describe the repository you have — not the repository an AI expects to exist.
README From Reality generates polished developer documentation from the codebase itself: manifests, scripts, config, entry points, environment examples, tests, docs, CI, and deployment clues.
The result is useful without pretending unknown things are known.
What it does
-
Builds a fact map from the actual repository before writing.
-
Detects the real package manager, scripts, configuration, entry points, test setup, and environment requirements.
-
Separates verified repository facts from reasonable inference and missing owner knowledge.
-
Refuses to invent installation commands, environment variables, features, APIs, integrations, or deployment methods just to make the README look complete.
-
Uses the repository’s existing terminology instead of replacing it with generic AI marketing copy.
-
Produces a complete, copy-ready README with only the sections the project actually supports.
-
Adds an evidence appendix showing what was verified, inferred, or still missing.
-
Lists commands that should be run before publication when execution evidence is not available.
-
Defines the repository changes that should trigger a future documentation update.
Use it when
- a repository has no README
- the README is stale or incomplete
- AI-generated documentation contains suspicious commands or claims
- onboarding instructions need to match the actual codebase
- setup steps, environment variables, or scripts need verification
- you want a professional README without hallucinated features
What you get
- Complete README draft
- Verified repository facts
- Clearly marked inferences
- Missing information that needs owner input
- Commands to verify before publishing
- Documentation update triggers
Why it is different
Most README generators optimize for completeness and polish. This one optimizes for usefulness without losing contact with the repository.
If the codebase does not support a claim, the skill does not quietly fill the gap.
Known limitations
A repository cannot reveal business positioning, private deployment procedures, undocumented external dependencies, or decisions that live only in team knowledge. Those gaps are surfaced explicitly instead of guessed.
Designed as a focused developer utility, not a general-purpose consultant. It produces a concrete artifact and explicitly separates verified evidence from assumptions when the available context is incomplete.
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 any payment.
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
2 uses
Times used to date
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 2 months ago
- Free with an Agensi account
Requires
Permissions
No special permissions declared or detected