Summary
Add clearer guardrails to CLAUDE.md so AI coding agents only include example or fictional information in comments, tests, documentation, and similar non-runtime artifacts.
Motivation
Real domains, customer identifiers, credentials, configuration details, or other potentially sensitive values can accidentally be introduced in comments, tests, docs, or example snippets. CLAUDE.md should explicitly tell agents to avoid creating or committing that kind of non-example information.
Acceptance criteria
- CLAUDE.md tells agents to use only example/fictional information in comments, tests, docs, examples, and similar materials.
- CLAUDE.md tells agents not to commit potentially sensitive real-world information in those contexts.
- The guidance complements the existing secret and
.env safeguards.
Summary
Add clearer guardrails to CLAUDE.md so AI coding agents only include example or fictional information in comments, tests, documentation, and similar non-runtime artifacts.
Motivation
Real domains, customer identifiers, credentials, configuration details, or other potentially sensitive values can accidentally be introduced in comments, tests, docs, or example snippets. CLAUDE.md should explicitly tell agents to avoid creating or committing that kind of non-example information.
Acceptance criteria
.envsafeguards.