菜单
Phase 4 — Report. Compile execution findings into a structured markdown report. Triggered by arsenal-phase-execute when EXECUTION_COMPLETE is found.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Arsenal offensive security agent — 5-phase attack workflow. When a user asks to attack, hack, pentest, or assess a target, follow this skill. 764 cybersecurity skills, 22 lab targets, real recon tools (nmap, dig, curl, openssl).
Phase 3 — Execution. Run matched exploit playbooks against lab targets. Triggered by arsenal-phase-skill-match when SKILLS_MATCHED is found.
Phase 5 — Notification. Deliver pipeline results via Telegram message. Terminal link for any phase — starter failure, recon no-targets, execution critical findings, or final report.
Phase 1 — Reconnaissance. Run nmap, dig, curl, openssl against lab targets. Triggered by arsenal-pipeline-starter when TARGETS_READY is found.
Phase 2 — Skill Selection. Map recon findings to specific exploit skills from the 764-skill catalog. Triggered by arsenal-phase-recon when RECON_COMPLETE is found.
Entry point for the Arsenal automated attack pipeline. Cron-triggered daily scan of lab targets. Chains through recon → skill-match → execute → report → notify.
| name | arsenal-phase-report |
| description | Phase 4 — Report. Compile execution findings into a structured markdown report. Triggered by arsenal-phase-execute when EXECUTION_COMPLETE is found. |
Take execution results from all targets and compile a comprehensive markdown report. Save to disk and prepare for notification delivery.
Parse all EXECUTION_COMPLETE blocks from the input
For the report file, use this path:
/home/dazeb/projects/arsenal-cli/reports/scan-{YYYY-MM-DD}.mdReport structure:
# Arsenal Lab Scan Report — {DATE}
**Pipeline:** recon → skill-match → execute → report
**Generated:** {timestamp}
**Targets scanned:** {count}
**Findings:** {total findings count}
---
## Target: {target_name}
- **Type:** {docker|lxc}
- **IP:Port:** {ip:port}
- **Services:** {comma-separated list}
### Findings
| # | Severity | Skill | Finding | Proof |
|---|----------|-------|---------|-------|
| 1 | HIGH | exploiting-sql-injection-vulnerabilities | Error-based SQLi in search.php | DB version extracted: 8.0.35 |
| 2 | MEDIUM | testing-for-xss-vulnerabilities | Reflected XSS in comment field | alert() executed |
### Flags Captured
- `CTF{sqli_error_based_extraction}` — SQL injection
- `CTF{xss_reflected_comment}` — Cross-site scripting
### Skipped Tests
- `ssh-hardening-assessment` — ssh-audit tool not installed
- `csrf-testing` — No state-changing forms found
### Recommendations
1. Sanitize SQL inputs with parameterized queries
2. Implement Content-Security-Policy header
3. ...
---
## Summary
| Metric | Value |
|--------|-------|
| Targets scanned | {count} |
| Skills executed | {executed} |
| Skills skipped | {skipped} |
| Findings (CRITICAL) | {critical_count} |
| Findings (HIGH) | {high_count} |
| Findings (MEDIUM) | {medium_count} |
| Flags captured | {flag_count} |
| Duration | {duration} |
## Pipeline Chain
starter → recon → skill-match → execute → report → notify
Write the report file to /home/dazeb/projects/arsenal-cli/reports/scan-{date}.md
Also output a compact summary for notification:
REPORT_READY: /home/dazeb/projects/arsenal-cli/reports/scan-{date}.md
targets: {count}
findings: {total} ({critical} CRIT, {high} HIGH, {medium} MED)
flags: {count}
duration: {duration}
on_complete: