vincent-credentials
von glitch003
AI & LLMsKurzbeschreibung
Use this skill to securely store credentials (API keys, passwords, tokens) and write them to .env files.
Beschreibung
Was dieser Skill macht
Manages API keys, passwords, OAuth tokens, and SSH keys for agents without exposing values to the conversation context. The agent creates a secret via the Vincent CLI, the user sets the value through the heyvincent.ai dashboard, and the CLI writes it directly to a .env file. The credential value never appears in stdout or the agent's context window.
Credentials never enter the agent's context or stdout, unlike pasting secrets directly into chat where they become part of the conversation history.
Wann du ihn einsetzt
- Storing a third-party API key for an app an agent is building
- Writing database username and password to .env without pasting them in chat
- Managing OAuth tokens for a service the agent needs to integrate
- Setting up SSH keys for a deployment pipeline being configured by an agent
- Recovering credential access after an agent loses its stored API key
Beispiel-Workflow
User asks: Store my Acme service credentials securely so the agent can use them
- 1
Check for existing credential keys with
secret list --type CREDENTIALS - 2
Create a new CREDENTIALS secret with
secret create --type CREDENTIALS --memo 'Acme credentials' - 3
Share the returned claimUrl with the user so they can set the value via the heyvincent.ai dashboard
- 4
Once the user confirms the value is set, run
secret env --key-id <ID> --env-var ACME_USERNAME --field usernameand repeat for password - 5
Confirm .env was updated using the returned confirmation JSON
.env file contains ACME_USERNAME and ACME_PASSWORD without the values ever appearing in the agent conversation
Voraussetzungen
- heyvincent.ai account (to claim secrets and set values via the dashboard)
Installation
clawhub install glitch003/vincent-credentialsFühre den CLI-Befehl aus oder füge den Prompt in Claude Code, Cursor, Codex oder deinen Coding-Agenten ein.