# JC PenTest Harness

Evidence-first, human-in-the-loop web cockpit for the Jean-Claude multi-model pentest harness.

Production URL: https://pentest.platphormnews.com
Updated: 2026-07-20T00:00:00.000Z

## Responsible Use
- Authorized testing only.
- Only test systems you own or have explicit written permission to assess.
- The operator is responsible for target authorization and scope.
- No anonymous active testing against arbitrary targets.
- No denial-of-service testing unless explicitly scoped and separately enabled.
- No destructive testing.
- No persistence, malware, credential theft, exfiltration, or public exploitation.
- No social engineering.
- No attacks against third-party infrastructure outside the approved scope.
- No remediation, PR creation, or branch push actions.
- Findings are assessment outputs, not automatic fixes.
- Human review is required before real execution and before publishing reports.
- Private findings and artifacts remain protected.

## Run Boundary
- Dry Run is the default and validates scope, config, model routing, prompts, and expected runner command without active target calls.
- Real Run requires PLATPHORM_API_KEY or authenticated operator context, accepted responsible-use policy, allowed hosts, explicit scope, and typed AUTHORIZED confirmation.
- Real Run starts bounded passive recon plus AI Gateway Hunt/Validate execution when Gateway credentials are present; otherwise it records an operator checkpoint instead of fake success.

## Discovery
- OpenAPI: https://pentest.platphormnews.com/openapi.yaml
- MCP: https://pentest.platphormnews.com/api/mcp
- Trust policy: https://pentest.platphormnews.com/.well-known/trust.json

## Models
- fable5: Fable 5 (anthropic, Primary Hunt + Validate)
- opus48: Claude Opus 4.8 (anthropic, Scorer of record + gapfill)
- gpt55: GPT-5.5 (openai, Cross-validation)

## Known Gaps
- Consolidated findings.jsonl is emitted by the CLI at both the per-model and top-level run directories.
- scoring.js implements priority scoring; web real-runs also derive scorecards from actual persisted findings.
- llm-router.route() exists, but the CLI runner scaffolds prompts and pauses for supervised invocation.
- Trust-boundary checks are preflight warnings for repo config and must be enforced by operator review and sandbox/job policy for active extensions.
- Sandbox and BrowserOps previews are public-safe. Protected handoffs only report success after a downstream response confirms receipt.

## Capabilities
- Jean-Claude boot and setup wizard: Setup wizard and run planner for scoped dry runs and protected real runs. (live_surface, /runs/new)
- Dry-run and real-run selection with dry-run default: Dry-run default, real-run gate, budget controls, and authorization confirmation. (live_surface, /runs/new)
- Execution modes: api-multi-model, local-client, local-model, multi-mode: Documented execution modes from the original harness configuration. (source_backed, /docs/modes)
- Model roster: fable5, opus48, gpt55: Configured model roster, provider selection, fallback models, pricing inputs, and stage role mapping. (source_backed, /models)
- Key resolution status for env, runtime keys, key files, AWS SSM, and Vercel AI Gateway: Presence-only key status for server envs and browser-session runtime keys. (live_surface, /settings)
- Target base URL, repositories, trust boundaries, and allowed-host validation: Target base URL, repositories, trusted hosts, and responsible-use scope controls. (live_surface, /runs/new)
- RECON, HUNT, VALIDATE, GAPFILL, TRACE, REPORT, SCORECARD pipeline timeline: Pipeline timeline and run history with persisted stage, log, finding, and scorecard evidence. (live_data, /runs)
- 44 attack-class taxonomy across 9 harness frameworks: Canonical Jean-Claude attack-class matrix, framework mapping, and latest class-level execution evidence. (source_backed, /matrix)
- Finding normalization, schema validation, run linkage, and evidence drilldown: Finding explorer with run linkage, target context, validation rationale, remediation text, and evidence boundaries. (live_data, /findings)
- Priority scoring, persisted scorecards, and comparative-scorecard gap disclosure: Priority scoring method, persisted scorecard rows, and explicit rubric boundary for evidence-backed findings. (live_data, /scorecard)
- Public-safe report packages and report ledger: Report packages generated from persisted runs, findings, scorecards, class evidence, and sanitized logs. (live_data, /reports)
- Prompt packages, class evidence grids, scorecards, findings summaries, and artifact inventory: Public-safe artifact inventory for run outputs, report package routes, class evidence, and downstream handoff artifacts. (live_data, /artifacts)
- Event outbox, audit trail, and live run progress refresh: Durable lifecycle event outbox, structured audit rows, canonical event catalog, and live refresh for in-flight runs. (live_data, /activity)
- Responsible-use audit and protected operator actions: Responsible-use audit policy and protected operator action boundary. (source_backed, /docs/responsible-use)
- Runtime browser-session credential entry for protected setup and handoff operator auth: Browser-session runtime credential entry for explicit operator auth and missing runtime dependencies. (live_surface, /settings)
- PlatPhorm Sandbox handoff preview and protected receive-handoff delivery: Sandbox handoff preview and protected receive-handoff delivery controls. (live_surface, /integrations)
- PlatPhorm BrowserOps handoff preview and protected receive-handoff delivery: BrowserOps handoff preview and protected receive-handoff delivery controls. (live_surface, /integrations)
- MCP, OpenAPI, llms, sitemap, RSS, robots, and well-known discovery: Discovery routes, API docs, MCP metadata, sitemap, RSS, robots, and well-known policy files. (generated, /docs)

## Public Routes
- /
- /runs
- /runs/new
- /findings
- /matrix
- /models
- /scorecard
- /reports
- /artifacts
- /activity
- /integrations
- /settings
- /known-gaps
- /registry
- /app/setup
- /app/setup/responsible-use
- /docs
- /docs/getting-started
- /docs/safety
- /docs/responsible-use
- /docs/modes
- /docs/pipeline
- /docs/findings
- /docs/integrations
- /api/health
- /api/v1/health
- /api/coverage
- /api/events
- /api/findings
- /api/stats
- /api/v1/setup/status
- /api/v1/responsible-use
- /api/v1/models
- /api/v1/models/status
- /api/v1/integrations/status
- /api/v1/integrations/platphorm/registry
- /api/v1/integrations/platphorm/discovery
- /api/v1/runtime/credentials
- /api/v1/reports
- /api/v1/events
- /api/v1/webhooks
- /api/docs
- /api/mcp
- /openapi.yaml
- /openapi.json
- /llms.txt
- /llms-full.txt
- /llms-index.json
- /humans.txt
- /robots.txt
- /sitemap.xml
- /sitemap-index.xml
- /sitemap-full.xml
- /rss.xml
- /feed.xml
- /manifest.webmanifest
- /.well-known/mcp.json
- /.well-known/agents.json
- /.well-known/ai-plugin.json
- /.well-known/agent-policy.json
- /.well-known/ai-policy.json
- /.well-known/security.txt
- /.well-known/trust.json

## Protected Routes
- /api/v1/runs custom dry-run scopes
- /api/v1/runs/real-run
- /api/v1/runs/:runId/rerun
- /api/v1/runs/:runId/(pause|resume|cancel)
- /api/v1/sandbox/jobs
- /api/v1/integrations/sandbox/handoff
- /api/v1/integrations/browserops/handoff
- /api/v1/integrations/*
- /api/v1/webhooks/*
- /api/mcp protected tools

## Attack Classes
- injection: Injection (A05:2025 / AML.T0051)
- broken-access-control: Broken Access Control (A01:2025)
- cryptographic-failures: Cryptographic Failures (A04:2025)
- security-misconfiguration: Security Misconfiguration (A02:2025)
- supply-chain-failures: Supply Chain Failures (A03:2025)
- auth-failures: Authentication Failures (A07:2025)
- ssrf: Server-Side Request Forgery (A01:2025 sub-class)
- secrets-in-code: Secrets in Code (cross-cutting)
- iac-privilege-escalation: IaC Privilege Escalation (cross-cutting)
- header-csp-hygiene: Header / CSP Hygiene (A02:2025)
- exceptional-conditions: Mishandling of Exceptional Conditions (A10:2025 (NEW))
- mcp-tool-poisoning: MCP Tool Poisoning (MCP / Agentic / AML.T0051)
- mcp-rug-pull: MCP Rug Pull (OWASP MCP Top 10)
- mcp-tool-shadowing: MCP Tool Shadowing (OWASP MCP Top 10)
- mcp-indirect-injection: MCP Indirect Injection (MCP / Agentic / AML.T0051)
- mcp-session-hijacking: MCP Session Hijacking (OWASP MCP Top 10)
- mcp-dns-rebinding: MCP DNS Rebinding (CVE-2025-9611)
- mcp-oauth-rce: MCP OAuth RCE (OWASP MCP Top 10)
- mcp-resource-exhaustion: MCP Resource Exhaustion (OWASP MCP Top 10)
- rag-retrieval-poisoning: RAG Retrieval Poisoning (USENIX Sec 2025)
- rag-embedding-inversion: RAG Embedding Inversion (VectuTex)
- rag-tenant-boundary-violation: RAG Tenant Boundary Violation (USENIX Sec 2025)
- rag-indirect-injection: RAG Indirect Injection (USENIX Sec 2025)
- agent-goal-hijacking: Agent Goal Hijacking (OWASP Agentic:2026)
- insecure-tool-execution: Insecure Tool Execution (OWASP Agentic:2026)
- memory-poisoning: Memory Poisoning (OWASP Agentic:2026 / AML.T0080)
- excessive-agency: Excessive Agency (OWASP Agentic:2026)
- agent-identity-abuse: Agent Identity Abuse (OWASP Agentic:2026)
- insecure-inter-agent-comms: Insecure Inter-Agent Comms (OWASP Agentic:2026)
- unbounded-autonomy: Unbounded Autonomy (OWASP Agentic:2026)
- mcp-escape-to-host: MCP Escape to Host (ATLAS:2026)
- emergent-offensive-reasoning: Emergent Offensive Reasoning (Cequence AZT v3.0)
- api-bola: API Broken Object Level Authz (API1:2023)
- api-bfla: API Broken Function Level Authz (API5:2023)
- api-mass-assignment: API Mass Assignment (API3:2023 / BLA3:2025)
- business-logic-abuse: Business Logic Abuse (OWASP BLA Top 10:2025)
- non-human-identity-abuse: Non-Human Identity Abuse (NIST SP 800-207 / CISA ZTMM)
- container-runtime-escape: Container Runtime Escape (MITRE ATT&CK T1611)
- k8s-rbac-lateral-movement: K8s RBAC Lateral Movement (MITRE ATT&CK T1078.004)
- cicd-pipeline-poisoning: CI/CD Pipeline Poisoning (CISA CSI CI/CD / SLSA)
- api-zombie-shadow-endpoint: Zombie / Shadow Endpoint (API9:2023)
- websocket-abuse: WebSocket Abuse (CVE-2026-24696 class)
- llm-model-weight-backdoor: LLM Model Weight Backdoor (NIST AI 100-2:2025 / AML.T0020)
- llm-inference-data-exfil: LLM Inference Data Exfil (AML.T0024 + AML.T0086)