Rebuild Vibeflow as an n8n safety gate

This commit is contained in:
Felipe Domingues 2026-07-22 11:43:41 -03:00
parent 79fc67166a
commit febec32131
101 changed files with 2405 additions and 3557 deletions

View file

@ -1,44 +1,15 @@
# Security Policy
# Security policy
## Reporting
Use GitHub private vulnerability reporting for this repository. Do not open a public issue containing exploit details, credentials, or customer workflow data.
## Supported versions
Security fixes target the latest tagged release.
## Scope
This repository contains prompts, conventions, examples, and packaging guidance for agent-driven workflow creation in n8n.
Vibeflow reads local JSON and optionally writes a caller-selected report. It does not contact n8n, execute workflows, collect telemetry, or upload workflow content.
It does not store production secrets by design.
## Reporting a vulnerability
If you discover a security issue related to:
- secret handling guidance,
- destructive workflow defaults,
- unsafe recipe recommendations,
- risky prompt behavior,
please report it privately before opening a public issue.
Use a private contact method for the maintainer when available.
## Security expectations for contributors
Contributors should avoid introducing guidance that:
- assumes access to credentials that may not exist,
- performs destructive actions without explicit user intent,
- sends external communications without confirmation,
- hides compliance-sensitive assumptions,
- suggests storing plaintext secrets in repo files.
## Safe defaults
The skill should prefer:
- placeholders over fake credentials,
- explicit assumptions over silent guesses,
- confirmation for destructive or externally visible actions,
- human-readable reports for manual review.
## Out of scope
This repository does not guarantee:
- security of any third-party MCP server,
- security of any n8n deployment,
- security of a user's local machine,
- correctness of external vendor SDKs or CLIs.
The secret rule is a defensive heuristic, not a replacement for repository secret scanning. A clean report is not proof of runtime security.