Skip to main content

/jaan-to:detect-dev

Repo engineering audit with machine-parseable findings and OpenSSF-style scoring.


What It Does

Performs a comprehensive engineering audit of the repository, producing 9 structured markdown reports covering stack, architecture, standards, testing, CI/CD, deployment, security, observability, and risks. Every finding is evidence-backed with SARIF-like locations and confidence scoring.

Scans manifest files (package.json, pyproject.toml, go.mod, Cargo.toml, etc.), Docker/compose configurations, CI/CD pipelines, git metadata, infrastructure-as-code, and project structure across 11+ language ecosystems.


Usage

/jaan-to:detect-dev
ArgumentRequiredDescription
repoNoTarget repository (defaults to current)

Output

FileContent
docs/current/dev/stack.mdTech stack with version evidence
docs/current/dev/architecture.mdArchitecture patterns and data flow
docs/current/dev/standards.mdCoding standards and conventions
docs/current/dev/testing.mdTest coverage and strategy
docs/current/dev/cicd.mdCI/CD pipelines and security
docs/current/dev/deployment.mdDeployment patterns
docs/current/dev/security.mdSecurity posture and findings (OWASP mapping)
docs/current/dev/observability.mdLogging, metrics, tracing
docs/current/dev/risks.mdTechnical risks and debt

Each file includes standardized YAML frontmatter + Findings blocks (ID/severity/confidence/evidence).


What It Scans

LayerConfidenceSources
Config files95-100%package.json, pyproject.toml, go.mod, Cargo.toml, Gemfile, composer.json, pom.xml, pubspec.yaml, mix.exs, Package.swift, *.csproj
Docker & databases90-95%docker-compose.yml, Dockerfile
CI/CD & testing90-95%GitHub Actions, GitLab CI, CircleCI, Jenkins, test configs, linters
Git & integrations95%git remote, CODEOWNERS, PR templates, Renovate/Dependabot
Infrastructure60-80%Terraform, serverless.yml, Vercel/Netlify/Fly configs, Kubernetes/Helm
Project structure60-80%Directory layout, monorepo signals

Key Points

  • Evidence IDs use namespace E-DEV-NNN (prevents collisions in pack-detect aggregation)
  • 4-level confidence: Confirmed (0.95-1.00) / Firm (0.80-0.94) / Tentative (0.50-0.79) / Uncertain (0.20-0.49)
  • Diataxis-style sections: Executive Summary → Scope/Methodology → Findings → Recommendations → Appendices
  • Frontmatter includes findings_summary buckets + overall_score (0-10, OpenSSF-style) + lifecycle_phase (CycloneDX)
  • CI/CD security checks: secrets boundaries, runner trust (self-hosted), permissions (write-all), action pinning (SHA vs @main), SLSA provenance
  • Overall score formula: 10 - (critical*2.0 + high*1.0 + medium*0.4 + low*0.1) / max(total_findings, 1)
  • Uses git tools for history analysis (git log, git remote, git show)

Back to Detect Skills | Back to All Skills