code-securitylisted
Install: claude install-skill Sheldon-92/TAD
**CONSUMES**: User code/repo + scan target URL (DAST) + IaC files + optional existing scan reports
**PRODUCES**: Applied security judgment rules + SARIF scan output + prioritized triage plan (P0-P3) + remediation actions + CI/CD gate configs
# Code Security Capability Pack
<!-- Verified against Semgrep v1.176.0 / Nuclei v3.11.1 / Gitleaks 8.30.1 / TruffleHog v3.97.4 / Checkov 3.3.16 / OSV-Scanner v2.5.1 / Grype v0.118.0 / Trivy v0.74.0 docs on 2026-09-11 via docs (CLIs ABSENT on impl host). Subcommands re-checked; only TruffleHog `--only-verified` → `--results=verified` renamed (see secret-detection-rules.md). -->
**Version**: 0.1.0
**Compatibility**: Claude Code (Phase 1); Codex / Cursor / Gemini in Phase 3
**License**: Apache 2.0
---
## What This Pack Does
AI agents set up security scanning by installing one tool and running it with default settings. They dump raw scanner output to the developer and call it done. They run SAST without DAST, missing runtime vulnerabilities entirely. They treat every CRITICAL CVSS as equally urgent, ignoring reachability and exploit probability. They detect secrets but clean up code before rotating the credential — leaving the attacker with a working key.
This pack embeds the judgment rules that application security engineers apply automatically — rules from Semgrep, Nuclei, Gitleaks, TruffleHog, Checkov, and vulnerability triage practices grounded in OWASP and real-world pipeline architecture.
**Pack = security judgment. Your workflow