Feature Roadmap

Current State (v0.0.1)

The extension already provides a strong foundation:

• Markdown-driven target/credential management with YAML blocks
• Environment variable export to all terminals
• CodeLens for shell commands, HTTP requests, YAML blocks
• msfvenom, hashcat, scanner integrations
• Terminal profiles (msfconsole, meterpreter, netcat, web delivery)
• Terminal recorder (command/output logging)
• Foam integration for knowledge graph + Tarjan SCC report generation
• Snippets: BloodHound, GTFOBins, LOLBAS
• CyberChef decoder integration

---

Phase 1: Foundation Hardening (v0.1.0)

Goal: Fix bugs, add tests, prepare for AI integration.

1.1 Code Quality Fixes

• [x] Fix missing await on Foam activation
• [x] Make Foam() a static method
• [x] Add error boundaries in activation
• [x] Convert let to const throughout
• [x] Fix duplicate is_dc assignment
• [x] State getter caching
• [x] defaultCollects lazy loading
• [ ] Enable strict TSConfig options

1.2 Testing

• [x] Set up test infrastructure with fixtures
• [x] Unit tests for core/domain/ (Host, UserCredential)
• [x] Unit tests for core/markdown/ (fencedBlocks, yamlBlocks)
• [x] Unit tests for core/env/ (collects, envVarSafer, mergeCollects)
• [ ] Integration tests for target sync
• [ ] E2E activation test
• [ ] CI test pipeline

1.3 Developer Experience

• [x] Add AGENTS.md for AI coding assistants
• [ ] Add .editorconfig
• [ ] Rewrite Python generators in TypeScript
• [ ] Document all environment variables in README

---

Phase 2: AI Integration (v0.2.0)

Goal: Add AI assistant and external AI control.

2.1 Copilot Chat Participant

• [x] Register @weapon Chat Participant
• [x] Implement /analyze command (analyze tool output)
• [x] Implement /suggest command (suggest next steps)
• [x] Implement /generate command (generate commands)
• [x] Implement /report command (engagement summary)
• [x] Implement /explain command (explain concepts)
• [x] Build system prompt with engagement context
• [x] Add credential redaction in all LLM contexts
• [ ] Add inline action buttons (run command, create note)

2.2 MCP Server

• [x] Implement embedded MCP server with Streamable HTTP transport
• [ ] Resources: hosts, users, env-vars, terminal-logs, notes
• [x] Read tools: get_targets, get_credentials, get_graph, list_findings, get_finding
• [x] Write tools: create_finding, update_finding_frontmatter
• [x] Formatted output tools: get_hosts_formatted, get_credentials_formatted
• [x] Terminal tools: list_terminals, read_terminal, send_to_terminal, create_terminal
• [ ] Not yet implemented: search_notes, get_attack_graph, switch_target, switch_user, run_command, run_scanner, generate_report, decode_text
• [ ] Prompt templates: analyze-output, suggest-next-steps, privesc-check
• [ ] MCP Inspector testing
• [ ] Documentation for Claude Code, Cursor, VS Code MCP config

2.3 AI Safety

• [ ] Credential redaction in all AI-facing interfaces
• [ ] Command validation for AI-initiated execution
• [ ] AI action audit log
• [ ] weaponized.ai.redactCredentials setting
• [ ] Command queue with user approval for MCP write tools

---

Phase 3: Enhanced Workflows (v0.3.0)

Goal: Smart automation and data import.

3.1 Auto-Import Scan Results

• [ ] File watcher for *.xml (nmap XML output)
• [ ] Nmap XML parser → auto-create host/service notes
• [ ] File watcher for *.json (various tool output)
• [ ] Import wizard: show preview, let user confirm before creating notes
• [ ] Support formats: nmap, masscan, nuclei JSON, feroxbuster JSON

3.2 Smart Command Suggestions

• [ ] Context-aware command palette: based on current host OS, services
• [ ] "What to try next" sidebar panel
• [ ] Command history with success/failure tracking
• [ ] Integration with terminal recorder: parse output → suggest follow-ups

3.3 Credential Vault

• [ ] Encrypt credentials at rest in workspaceState
• [ ] Master password on workspace open (or keychain integration)
• [ ] Credential rotation tracking (when was this password obtained?)
• [ ] weaponized.vault.enabled setting

3.4 Multi-Engagement Support

• [ ] Engagement profiles (different state per engagement)
• [ ] Quick-switch between engagements
• [ ] Engagement archiving (export all notes + state as ZIP)

---

Phase 4: Visualization & Reporting (v0.4.0)

Goal: Rich visual analysis tools.

4.1 Attack Graph Webview

• [ ] Interactive webview panel for the attack graph (not just Mermaid in markdown)
• [ ] Clickable nodes → open corresponding Foam note
• [ ] Color-coded by node type (host, user, service, finding)
• [ ] Filter by severity, host, technique

4.2 Timeline View

• [ ] Webview panel showing chronological timeline of:
  • Terminal commands executed
  • Hosts discovered
  • Credentials obtained
  • Findings documented
• [ ] Filterable by host, user, time range
• [ ] Export as markdown for report appendix

4.3 Report Templates

• [ ] Multiple report formats: markdown, HTML, PDF (via pandoc)
• [ ] Executive summary template
• [ ] Technical findings template (per-finding: description, evidence, remediation)
• [ ] CVSS scoring integration
• [ ] Report diff: compare two engagement snapshots

4.4 Dashboard

• [ ] Status bar items: current host, current user, # findings
• [ ] Tree view: engagement overview (hosts → services → findings)
• [ ] Quick stats: hosts owned, users compromised, findings by severity

---

Phase 5: Collaboration & Ecosystem (v0.5.0)

Goal: Team features and tool ecosystem.

5.1 Team Collaboration

• [ ] Shared engagement state via Git (already markdown-based!)
• [ ] Conflict resolution for simultaneous edits
• [ ] Activity feed: what did other team members discover?
• [ ] Shared terminal log aggregation

5.2 Tool Ecosystem

• [ ] Plugin API for custom tool integrations
• [ ] Community snippet packs (beyond BloodHound/GTFOBins/LOLBAS)
• [ ] Custom scanner definitions (beyond weaponized.scanners config)
• [ ] Integration with:
  • Burp Suite (import/export)
  • BloodHound CE (direct API)
  • Cobalt Strike (team server listener status)
  • Sliver C2
  • Havoc C2

5.3 AI Enhancements

• [ ] Local LLM support (ollama) for air-gapped engagements
• [ ] AI-powered finding deduplication
• [ ] Auto-remediation suggestions per finding
• [ ] AI report writing assistant (draft → review → finalize)
• [ ] VS Code vscode.lm.registerTool() when API stabilizes (replaces standalone MCP)

---

Feature Priority Matrix

Feature	User Value	Effort	Dependencies	Phase
Code quality fixes	Medium	Low	None	1
Unit tests for core/	High	Low	None	1
Copilot Chat Participant	High	Medium	None	2
MCP Server (read-only)	High	Medium	None	2
MCP Server (tools)	High	Medium	MCP read-only	2
Nmap XML auto-import	High	Medium	None	3
Credential vault	High	Medium	None	3
Attack graph webview	Medium	High	Report generator	4
Timeline view	Medium	High	Terminal recorder	4
Local LLM support	Medium	High	AI service layer	5
Team collaboration	Medium	High	Git workflow	5
Tool ecosystem plugins	Low	Very High	Plugin API design	5

---

Non-Goals

These are explicitly out of scope:

• Building a C2 framework — use existing ones (Sliver, Cobalt Strike)
• Replacing Burp Suite — focus on the terminal/note-taking workflow
• Web-based UI — this is a VS Code extension, not a web app
• Automated exploitation — the AI assists, the human decides and acts
• Supporting non-VS Code editors — the MCP server enables AI tool integration, but the core experience is VS Code

---

Success Metrics

Metric	Current	Phase 1 Target	Phase 2 Target
Test coverage (core/)	~90%	90%	90%
Test coverage (overall)	0%	40%	60%
Known bugs	~5	0	0
AI commands available	5	0	5+
MCP tools available	13	0	10+
Scan result auto-import	0 formats	0	0
Supported scanner configs	9	9	12+