mirror of https://github.com/mukul975/Anthropic-Cybersecurity-Skills.git synced 2026-06-11 21:54:56 +03:00

Files

T

mukul975 27c6414ca5 Add folder anatomy (scripts/agent.py + references/api-reference.md) for 648 cybersecurity skills

Complete skill folder anatomy across all cybersecurity skills:
- scripts/agent.py: 80-150 line Python agents using real libraries (impacket,
  boto3, azure-mgmt-*, kubernetes, pefile, yara, scapy, shodan, stix2, etc.)
- references/api-reference.md: real API documentation with method signatures
- LICENSE: MIT license for all skill folders

2026-03-10 21:02:12 +01:00

2.8 KiB

Raw Blame History

API Reference: Threat Intelligence Report Generator Agent

Dependencies

Library	Version	Purpose
jinja2	>=3.1	Template rendering for report generation

CLI Usage

python scripts/agent.py \
  --type operational \
  --data /cases/intel_data.json \
  --output-dir /cases/reports/ \
  --output report_meta.json

Report Types

Type	Audience	Length	Frequency
strategic	C-suite, board, risk committee	1-3 pages	Monthly/Quarterly
operational	CISO, security directors, IR leads	3-8 pages	Weekly
tactical	SOC analysts, threat hunters	1-2 pages	Daily/as-needed
flash	All security staff	1 page max	Urgent/as-needed

Functions

`confidence_label(level) -> str`

Maps confidence levels to ICD 203 language: "high" -> "We assess with high confidence", "medium" -> "We assess", "low" -> "Evidence suggests".

`render_report(report_type, data) -> str`

Renders a Jinja2 template with the provided data dict. Sets defaults for date, org, tlp.

`validate_report_data(report_type, data) -> list`

Validates required fields per report type. Returns list of error strings.

`quality_check(rendered) -> list`

Checks rendered report for: minimum length, TLP marker presence, unqualified confidence statements.

`generate_report(report_type, data_path, output_dir) -> dict`

Full pipeline: load JSON data, validate, render template, run quality checks, save Markdown output.

TLP Levels

Level	Sharing Scope
RED	Named recipients only
AMBER+STRICT	Organization only
AMBER	Organization and trusted partners
GREEN	Community-wide (ISAC, sector peers)
CLEAR	Public distribution

Input Data Schema (Operational Example)

{
  "title": "APT29 Campaign Targeting Financial Sector",
  "tlp": "AMBER",
  "org": "Security Operations Center",
  "executive_summary": ["APT29 actively targeting financial institutions..."],
  "adversary": {
    "name": "APT29 / Cozy Bear",
    "motivation": "Espionage",
    "sophistication": "Advanced",
    "target_sectors": ["Financial", "Government"]
  },
  "ttps": [{"tactic": "Initial Access", "technique_id": "T1566.001", "name": "Spearphishing", "observed": "2025-03-01"}],
  "key_judgments": [{"confidence": "high", "statement": "APT29 will continue targeting...", "evidence": "..."}],
  "recommendations": [{"priority": "Critical", "description": "...", "owner": "SOC", "timeframe": "24h", "details": "..."}],
  "iocs": [{"type": "domain", "value": "evil[.]com", "context": "C2", "confidence": "high"}]
}

Output

The agent produces two files:

{type}_report_{date}.md - Rendered Markdown report with TLP headers
report_meta.json - Metadata including validation errors and quality issues

2.8 KiB Raw Blame History