deep-app-audit — Claude Code skill: audit your Frappe app
deep-app-audit is a Claude Code skill for auditing Frappe apps for security issues, logic errors, and customization defects. Parallel agents check separate code aspects in fresh contexts and compile one report.
no license 458 installs ★ 177 updated 2026-10-24
npx skills add frappe/skills --skill deep-app-audit -a claude-codeWhat deep-app-audit does
Deep-app-audit examines your app across three dimensions: security (authorization, authentication, injection vectors and more), logic correctness (typing, error handling, state), and customization quality (alignment with Frappe standards). Each dimension breaks into separate checks — up to 160 agents on a full run.
The skill orchestrates a distributed process: sets up a test Frappe site (optional), builds an entry-point inventory, launches parallel scans with Semgrep, then verifies each finding in an isolated agent context. Results go to files, so a stopped run resumes from where it left off. You need Python, a Frappe bench, and local build tools.
You stay in control of the run: pick which checks to run (only: S-A, Q-B05), decide whether to create a test site, and set agent parallelism. The command /deep-app-audit [path] [options] kicks off the process; the run_audit.py scripts do the heavy lifting.
What it helps with
- detect authorization and injection vulnerabilities across the codebase
- verify type safety and error handling throughout the application
- ensure customizations follow Frappe conventions and standards
- validate each finding in an independent, fresh context
- produce a detailed report with severity levels for triage
Example prompts for the agent
- Run
/deep-app-audit ./my_appto audit the entire application - Check security only:
/deep-app-audit ./my_app only: S- - Skip expensive checks:
/deep-app-audit ./my_app skip: S-L no checks jobs: 8
How to install deep-app-audit
1Skills CLI (Vercel)
npx skills add frappe/skills --skill deep-app-audit -a claude-code2Manually: put the files in place
~/.claude/skills/deep-app-audit/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/deep-app-audit/
1Skills CLI (Vercel)
npx skills add frappe/skills --skill deep-app-audit -a codex2Manually: put the files in place
.agents/skills/deep-app-audit/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add frappe/skills --skill deep-app-audit -a cursor2Manually: put the files in place
.cursor/skills/deep-app-audit/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install frappe/skills skills/deep-app-audit2Manually: put the files in place
.github/skills/deep-app-audit/SKILL.md1Skills CLI (Vercel)
npx skills add frappe/skills --skill deep-app-audit -a gemini-cli2Platform CLI
gemini skills install https://github.com/frappe/skills --path skills/deep-app-auditManually: put the files in place
.devin/skills/deep-app-audit/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add frappe/skills --skill deep-app-audit -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about deep-app-audit
How do I install deep-app-audit in Claude Code?
Run npx skills add frappe/skills --skill deep-app-audit -a claude-code, then open ~/.claude/skills/deep-app-audit/SKILL.md to connect the skill.
What does deep-app-audit check?
The skill scans for security issues (authorization, injection, XSS), logic errors (types, state, error handling), and adherence to Frappe standards in customizations. Each check runs as a separate agent.
Do I need to create a test site?
By default, the skill creates a temporary Frappe site to run checks. If your bench is unavailable or looks like production, use no site to analyze source code only; some checks will be skipped.