Oathe Security Badge

oathe / scan / reports / safe

Is better-auth/skills safe?

Yes. better-auth/skills is safe to install. Oathe's behavioral security audit gave the skills skill by better-auth a trust score of 91/100 with 5 findings, none critical or high. Report updated

https://github.com/better-auth/skills

91
SAFE

Is better-auth/skills safe to install?

The better-auth/skills repository is a legitimate, documentation-only plugin published by the official Better Auth organization. All SKILL.md files contain straightforward authentication scaffolding guidance with no prompt injection directives, no instructions to access sensitive files, no executable code, and no unexpected network behavior during clone. The only notable concern is a low-severity indirect injection surface via instructions to fetch better-auth.com/llms.txt at agent runtime, which would only become exploitable if that domain were compromised. Canary file accesses observed in monitoring are attributable to the monitoring infrastructure itself, not to any behavior triggered by the repository.

What security issues were found in better-auth/skills?

Category Scores

Prompt Injection 88/100 · 30%
Data Exfiltration 91/100 · 25%
Code Execution 97/100 · 20%
Clone Behavior 93/100 · 10%
Canary Integrity 95/100 · 10%
Behavioral Reasoning 90/100 · 5%

Findings (5)

LOW External documentation fetch instruction -12 ▶

Both best-practices/SKILL.md and create-auth/SKILL.md instruct the agent to fetch better-auth.com/llms.txt and 'follow the matching version index' when no MCP is available. This creates an indirect prompt injection surface: if the domain is ever compromised or served malicious content, the agent would incorporate that content into its decision-making without the user's awareness.

Evidence: best-practices/SKILL.md: 'Without MCP, start at [better-auth.com/llms.txt](https://better-auth.com/llms.txt) and follow the matching version index.'

LOW Broad filesystem scanning instructions -9 ▶

create-auth/SKILL.md instructs the agent to scan the project for framework configs, lockfiles, ORM schemas, existing auth library imports, and package manager artifacts. While appropriate for the skill's purpose, this exposes detailed project structure to the agent context and could surface sensitive file locations (e.g., .env paths, database connection strings in drizzle.config.ts) that then persist in context.

Evidence: create-auth/SKILL.md Phase 1 Step 1: 'Look for prisma/schema.prisma, drizzle.config.ts, package.json deps (pg, postgres, @neondatabase/serverless, mysql2, better-sqlite3, mongoose, mongodb). If drizzle.config.ts exists, read its dialect field...'

INFO Documentation-only repository 0 ▶

The repository contains exclusively markdown SKILL.md files and JSON metadata. No executable code, no npm scripts, no git hooks, no compiled artifacts. Installation risk from code execution is negligible.

Evidence: Filesystem diff shows only .md, .json, and .git metadata files. package.json search returned empty. .gitattributes search returned empty.

INFO Single expected network destination during clone 0 ▶

The only external TCP connection during the git clone was to 140.82.121.3:443, which is a known GitHub IP. No unexpected outbound connections, no DNS lookups to non-standard domains, no background processes.

Evidence: auditd SOCKADDR: saddr=020001BB8C5279030000000000000000 → saddr_fam=inet laddr=140.82.121.3 lport=443. Connection diff shows no new persistent connections after install.

INFO Canary file accesses attributed to monitoring infrastructure 0 ▶

Canary files (.env, .ssh/id_rsa, .aws/credentials, .npmrc, .docker/config.json, GCP ADC) were accessed at timestamps 1794816957 (pre-clone, monitoring setup) and 1794816979 (post-analysis, integrity verification). Neither event is causally linked to repository content, git hooks, or any exec event from the cloned repo. Monitoring system confirms all files intact.

Evidence: auditd PATH events at 1794816957.888 (pre-clone) and 1794816979.716 (post-analysis). Clone exec event at 1794816963.410. Monitoring result: '✅ All canary files intact'.

Should I install better-auth/skills?

Oathe's verdict for better-auth/skills is SAFE with a trust score of 91/100. Recommendation: Install.