agentleFS
jaskaranhundaljaskaranhundal/usap-skills/appsec-devsecops/CLAUDE.md

usap-skills / appsec-devsecops

CLAUDE.md4 starsChanged 2 months agoApache-2.0View on GitHub

This file is the authoritative domain guide for the appsec-devsecops/ directory. It governs how Claude and cs-* agents understand, navigate, and apply the skills in this domain. The AppSec & DevSecOps domain embeds security into every stage of the software development lifecycle (SDLC) — from requirements and design through code review, pipeline execution, artifact signing, and post-deployment monitoring. Skills in this domain operationalize shift-left security by placing automated gates at developer touchpoints (pre-commit, PR, CI/CD) and by providing structured analysis…

The file

We split this CLAUDE.md into the rules it gives an agent

6 rules. About 0 are common, 2 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Shift left without blocking velocity: Place the fastest, highest-signal gates earliest (pre-commit: <30 s, PR gate: <5 min).

    line 104
  2. Distinctive:

    Gate on severity, not volume: Only Critical and High findings block merge or deployment.

    line 106
  3. Very specific:

    Deduplicate before escalating: Run sast-dast-coordinator before routing findings to developers.

    line 112
  4. Very specific:

    Threat model before you build: Use secure-sdlc to assess security requirements at the design phase for any change touching authentication, …

    line 116
  5. Very specific:

    Simulate to validate: Run supply-chain-simulation quarterly and after significant dependency tree changes to validate that detection …

    line 118
  6. Very specific:

    When a Critical finding is produced by any skill in this domain, route immediately to cs-security-analyst for triage.

    line 204

≈ “Common” is an estimate: at least 1 in 5 of the 27,641 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
CLAUDE.md

AppSec & DevSecOps Domain — CLAUDE.md

This file is the authoritative domain guide for the appsec-devsecops/ directory. It governs how Claude and cs-* agents understand, navigate, and apply the skills in this domain.


Purpose

The AppSec & DevSecOps domain embeds security into every stage of the software development lifecycle (SDLC) — from requirements and design through code review, pipeline execution, artifact signing, and post-deployment monitoring. Skills in this domain operationalize shift-left security by placing automated gates at developer touchpoints (pre-commit, PR, CI/CD) and by providing structured analysis of code, dependencies, and build pipelines.

Core coverage areas:

  • Secure SDLC — Security requirements, threat modeling, design review, and maturity assessment aligned to OWASP SAMM and BSIMM
  • SAST / DAST — Coordinating static and dynamic analysis results across multiple scanners; deduplicating and prioritizing findings
  • DevSecOps Pipeline — Assessing security gate completeness and correctness in CI/CD configurations
  • Build Integrity — Verifying artifact provenance, signature validity, and SLSA compliance to prevent build-time compromise
  • Supply Chain Risk — SBOM generation and analysis, malicious package detection, dependency risk scoring, and SLSA level assessment
  • AppSec Code Review — OWASP Top 10 and CWE-focused static analysis of pull requests used as a merge gate
  • Pipeline Security Scan — Active scanning of CI/CD YAML configurations for secrets exposure, SAST gaps, and insecure action pinning

The orchestrating agent for this domain is cs-devsecops-engineer.


Skills Catalog

SkillSlugPrimary ToolCoverage
Secure SDLCappsec-devsecops/secure-sdlcsecure-sdlc_tool.pyOWASP SAMM, BSIMM
SAST/DAST Coordinatorappsec-devsecops/sast-dast-coordinatorsast-dast-coordinator_tool.pyCWE Top 25, OWASP Top 10
DevSecOps Pipelineappsec-devsecops/devsecops-pipelinedevsecops-pipeline_tool.pyCI/CD security gates
Build Integrityappsec-devsecops/build-integritybuild-integrity_tool.pySLSA L1-L4, Sigstore
Supply Chain Riskappsec-devsecops/supply-chain-risksupply-chain-risk_tool.pySBOM, SLSA, Scorecard
Supply Chain Simulationappsec-devsecops/supply-chain-simulationsupply-chain-simulation_tool.pyAttack simulation
AppSec Code Reviewappsec-devsecops/appsec-code-reviewappsec-code-review_tool.pyOWASP Top 10, CWE
Pipeline Security Scanappsec-devsecops/pipeline-security-scanpipeline-security-scan_tool.pySecrets, SAST gaps
Security Requirements Reviewappsec-devsecops/security-requirements-reviewsecurity-requirements-review_tool.pyDesign doc intake, PRD/architecture/POA&M analysis

Each skill directory follows the USAP Agent Skills Standard v1 layout:

<skill-slug>/
├── SKILL.md
├── README.md
├── scripts/
│   └── <skill-slug>_tool.py
├── references/
├── assets/
└── expected_outputs/

Python Tools Reference

ToolPathPrimary Use CaseOutput
appsec-code-review_tool.pyappsec-code-review/scripts/PR security gate; OWASP Top 10 static analysisJSON findings with CWE mappings
pipeline-security-scan_tool.pypipeline-security-scan/scripts/Scan pipeline YAML for secrets, SAST gaps, pinning issuesJSON scan report
sast-dast-coordinator_tool.pysast-dast-coordinator/scripts/Normalize and deduplicate multi-scanner resultsDeduplicated finding set
supply-chain-risk_tool.pysupply-chain-risk/scripts/SBOM analysis, malicious package detection, license auditRisk-scored SBOM report
build-integrity_tool.pybuild-integrity/scripts/Artifact signature and SLSA provenance verificationVerification result + gate decision
devsecops-pipeline_tool.pydevsecops-pipeline/scripts/Assess security gate coverage in CI/CD pipelinePipeline security score
secure-sdlc_tool.pysecure-sdlc/scripts/SDLC maturity scoring and security requirements auditMaturity score (0-5) + gap list
supply-chain-simulation_tool.pysupply-chain-simulation/scripts/Simulate supply chain attack scenariosSimulation report
security-requirements-review_tool.pysecurity-requirements-review/scripts/Ingest design docs (PRD, arch, POA&M); extract security gaps; route findingsJSON output contract with document_metadata + design_analysis

Invoke all tools with --output json for machine-readable output compatible with downstream aggregation.


SDLC Phase Integration Matrix

SDLC PhaseSkills ActiveGatesArtifacts Produced
Requirements & Designsecure-sdlc, security-requirements-reviewThreat model sign-off required for high-risk features; document security review before design freezeThreat model, abuse case register, security requirements doc, design gap report
Development (pre-commit)appsec-code-review, pipeline-security-scanPre-commit hook: secrets scan + lint (<30 s)Pre-commit finding log
Pull Requestappsec-code-review, sast-dast-coordinator, supply-chain-riskBlock merge on Critical/High findingsPR security gate report, CWE-mapped findings
CI/CD Builddevsecops-pipeline, pipeline-security-scan, build-integrityFail pipeline on missing SAST gate or unsigned artifactPipeline scan report, artifact signature
Pre-Productionsast-dast-coordinator, supply-chain-risk, build-integrityDeployment gate: SLSA L2+ required; no Critical CVEsSBOM, SLSA provenance, pentest sign-off
Operationssupply-chain-risk, supply-chain-simulationContinuous dependency monitoring; quarterly simulationDependency diff report, simulation results

OWASP Top 10 Coverage (2021)

CategorySkillsDetection Method
A01 — Broken Access Controlappsec-code-review, sast-dast-coordinatorStatic pattern matching; DAST auth bypass probes
A02 — Cryptographic Failuresappsec-code-review, secure-sdlcAlgorithm allowlist check; TLS version scan
A03 — Injectionappsec-code-review, sast-dast-coordinatorParameterized query enforcement; DAST injection payloads
A04 — Insecure Designsecure-sdlcThreat model review; abuse case analysis
A05 — Security Misconfigurationpipeline-security-scan, devsecops-pipelinePipeline YAML audit; default credential checks
A06 — Vulnerable and Outdated Componentssupply-chain-risk, sast-dast-coordinatorSBOM diff against CVE database; SCA results coordination
A07 — Identification and Auth Failuresappsec-code-review, secure-sdlcSession management pattern review; JWT validation check
A08 — Software and Data Integrity Failuresbuild-integrity, supply-chain-riskArtifact signing verification; SLSA provenance check
A09 — Security Logging and Monitoring Failuressecure-sdlc, appsec-code-reviewLogging coverage review; sensitive data in logs pattern
A10 — Server-Side Request Forgeryappsec-code-review, sast-dast-coordinatorURL validation pattern check; DAST SSRF probes

Domain Best Practices

  1. Shift left without blocking velocity. Place the fastest, highest-signal gates earliest (pre-commit: <30 s, PR gate: <5 min). Reserve longer scans (DAST, full SCA) for CI/CD after commit, not before.
  2. Gate on severity, not volume. Only Critical and High findings block merge or deployment. Medium findings create tracked issues. Low and Informational findings are comments only. Tuning gates to this threshold keeps false-positive-driven bypasses below 15%.
  3. Every artifact must have a cryptographic chain of custody. From source commit hash to deployed container digest, there must be no unsigned gaps. Enforce SLSA L2 as the minimum for production deployments; target SLSA L3 for internet-facing services.
  4. SBOM is a first-class release artifact. Generate a CycloneDX or SPDX SBOM for every software release. Store it alongside the build provenance. Make it available to security and legal before release sign-off.
  5. Deduplicate before escalating. Run sast-dast-coordinator before routing findings to developers. A single SQL injection finding reported by three scanners must appear once, with the highest confidence and clearest remediation.
  6. Pin and verify all third-party CI actions. Pipeline configurations that reference mutable action tags (@main, @v2) are supply chain attack vectors. Require SHA pinning (@<full-commit-sha>) for all external actions and verify with pipeline-security-scan.
  7. Threat model before you build. Use secure-sdlc to assess security requirements at the design phase for any change touching authentication, authorization, cryptography, PII storage, or new external integrations. A missed threat model is a debt that compounds through every downstream gate.
  8. Simulate to validate. Run supply-chain-simulation quarterly and after significant dependency tree changes to validate that detection controls (SBOM diff, package hash verification, build anomaly detection) catch realistic supply chain attack scenarios before a real attacker does.

Workflow: PR Security Gate

This is the primary daily workflow. It executes on every pull request before merge is permitted.

Step 1 — Code Review (appsec-code-review)
  python appsec-code-review/scripts/appsec-code-review_tool.py --output json
  Gate: Block on Critical or High findings

Step 2 — Scanner Coordination (sast-dast-coordinator)
  python sast-dast-coordinator/scripts/sast-dast-coordinator_tool.py --output json
  Gate: Deduplicate and re-score; re-apply Critical/High block rule

Step 3 — Dependency Audit (supply-chain-risk)
  python supply-chain-risk/scripts/supply-chain-risk_tool.py --output json
  Gate: Block if new dependency has Critical CVE or is flagged malicious

Step 4 — Decision
  PASS  → Approve PR; route Medium/Low findings to tracking system
  BLOCK → Return findings to developer with CWE mappings and remediation steps
          Require security team sign-off for any override of High findings
          Require CISO sign-off for any override of Critical findings

Step 5 — Tracking
  Route all findings to findings-tracker for lifecycle and SLA management

Expected output: A structured PR security gate decision (pass/block) with severity-sorted findings, CWE mappings, and developer-actionable remediation guidance.


Workflow: Release Security Checklist

This workflow gates a software release before production deployment.

Step 1 — Pipeline Integrity Review
  python pipeline-security-scan/scripts/pipeline-security-scan_tool.py --output json
  python devsecops-pipeline/scripts/devsecops-pipeline_tool.py --output json
  Requirement: No secrets in pipeline env vars; all stages present

Step 2 — Build Artifact Verification
  python build-integrity/scripts/build-integrity_tool.py --output json
  Requirement: Signature valid; SLSA L2 minimum; provenance present

Step 3 — SBOM Generation and Risk Scoring
  python supply-chain-risk/scripts/supply-chain-risk_tool.py --output json
  Requirement: No Critical CVEs in SBOM; no malicious packages; license compliance confirmed

Step 4 — SDLC Maturity Check
  python secure-sdlc/scripts/secure-sdlc_tool.py --output json
  Requirement: Maturity score >= 2 (SAST + SCA in place)

Step 5 — Release Decision
  ALL PASS → Release approved; attach SBOM and provenance to release artifact
  ANY FAIL → Release blocked; create incident ticket; notify release manager

Artifacts produced: SBOM (CycloneDX or SPDX), build provenance (SLSA .intoto.jsonl),
                   pipeline scan report, release gate decision record

Cascade Intelligence (Cross-Skill Routing)

security-requirements-review ──► risk-threat-modeling (architecture docs)
                              ──► compliance-mapping (regulated data detected)
                              ──► appsec-code-review (PRD / requirements)
                              ──► pipeline-security-scan (pipeline/CI references)
                              ──► cs-security-analyst (critical gaps: no auth, hardcoded creds)

appsec-code-review ──► sast-dast-coordinator ──► findings-tracker
                                                 └──► secrets-exposure (hardcoded credential findings)

supply-chain-risk ──► build-integrity ──► incident-commander (Critical integrity failures)

devsecops-pipeline ──► pipeline-security-scan ──► secure-sdlc

supply-chain-simulation ──► red-team/ (validated attack scenarios)

When a Critical finding is produced by any skill in this domain, route immediately to cs-security-analyst for triage. Do not wait for the full gate pipeline to complete.


Related Domains

DomainDirectoryRelationship
Detectiondetection/AppSec findings that survive to production become detection use cases (e.g., SSRF patterns become WAF rules and SIEM detections)
Red Teamred-team/Supply chain simulation outputs feed red team attack scenarios; red team validates AppSec gate effectiveness
Risk & Compliancerisk-compliance/SBOM and SLSA provenance feed compliance evidence; OWASP SAMM maturity scores feed risk register
Responseresponse/Critical build integrity failures and confirmed supply chain compromises trigger incident response

Skill Level Reference

LevelMeaningSkills at this Level
L3Practitioner — structured analysis, defined outputsecure-sdlc, sast-dast-coordinator, devsecops-pipeline, build-integrity, supply-chain-risk, supply-chain-simulation, security-requirements-review
L4Expert — blocking gate authority, CWE mapping, high-confidence decisionsappsec-code-review, pipeline-security-scan

L4 skills have authority to produce block_required: true decisions. L3 skills produce findings and recommendations that feed into gate decisions made by L4 skills or the orchestrating agent.


Standards and Frameworks Referenced

StandardApplication in this Domain
OWASP SAMM 2.0Secure SDLC maturity model; secure-sdlc maturity scoring
BSIMM 14Observed security practices benchmark; gap analysis for secure-sdlc
OWASP Top 10 (2021)Primary vulnerability taxonomy for appsec-code-review and sast-dast-coordinator
CWE Top 25Secondary finding classification; CWE IDs attached to all findings
SLSA Framework (L1-L4)Build integrity and provenance requirements; build-integrity assessment
Sigstore / cosignArtifact signing mechanism referenced in build-integrity
SPDX 2.3 / CycloneDX 1.5SBOM format standards; supply-chain-risk output formats
OpenSSF ScorecardDependency health scoring; referenced in supply-chain-risk
NIST SP 800-218 (SSDF)Secure Software Development Framework; aligns with secure-sdlc phases

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.