diff --git a/docs/images/live-01-create-form.jpg b/docs/images/live-01-create-form.jpg index 3feb5a4..3ed2871 100644 Binary files a/docs/images/live-01-create-form.jpg and b/docs/images/live-01-create-form.jpg differ diff --git a/docs/live-campaign-walkthrough.md b/docs/live-campaign-walkthrough.md index 1194646..948e787 100644 --- a/docs/live-campaign-walkthrough.md +++ b/docs/live-campaign-walkthrough.md @@ -1,9 +1,10 @@ # Walkthrough: run a live campaign end to end This is a complete, copy-pasteable tutorial for setting up GoPhish and running a -full live phishing simulation in Pretexta — from zero to a running campaign with -AI follow-up. Every step here was executed against the real app; the screenshots -are from that run. +full live phishing simulation in Pretexta — from zero to a running **click +campaign** measured by **report rate**. (The optional AI-conversation layer is +covered in step 9.) Every step here was executed against the real app; the +screenshots are from that run. > Live mode sends real email. Do this first against **sandbox domains** > (`*.test`) and a local mail server, as shown here. Only point it at real @@ -22,6 +23,7 @@ Contents: 10. [Results and the kill switch](#10-results-and-the-kill-switch) 11. [Going to real mailboxes](#11-going-to-real-mailboxes) 12. [Troubleshooting](#12-troubleshooting) +13. [After the campaign — the rest of LIVE OPS](#13-after-the-campaign--the-rest-of-live-ops) --- @@ -164,29 +166,51 @@ For a real domain you would add it, then publish the shown DNS **TXT** record ## 7. Create a campaign -Click **+ New campaign**. You can **Start from a template** (built-in phishing +Click **Kampanye baru** (top-right). You can **Start from a template** (built-in pretexts: IT helpdesk, CEO wire, HR policy, delivery, tool install) to fill the -fields, then adjust — or fill the form yourself: +fields, then adjust — or fill the form yourself. ![New campaign form](images/live-01-create-form.jpg) +**First choose the campaign type.** The form opens on the two choices: + +- **Kampanye Klik** *(the default — start here)* — you send a lure and measure + behaviour: who clicks, and the headline metric, **who reports it**. No AI, no + conversation. This is the primary flow and the one this walkthrough runs. +- **Percakapan AI** *(advanced)* — an AI persona continues a reply thread toward + one approved objective (`reply`, `agree_to_share_otp`, …). Every reply is held + for human review (see step 9). Pick this only when you specifically want the + conversation layer; it needs IMAP + an LLM configured. + +For a click campaign, fill: + | Field | Example | |-------|---------| | Name | Q4 Helpdesk Phishing Drill | -| Objective | `agree_to_share_otp` | | Sender name / email | IT Helpdesk / `helpdesk@acme.test` (must be on a verified domain) | | Influence principles | authority, scarcity | | Subject | Action required: verify your mailbox | -| Body (HTML) | `

We detected unusual activity. Confirm within 2 hours: {{.URL}}

` | +| Body (HTML) | `

We detected unusual activity. Confirm here: verify now

` | | Recipients | `alice@acme.test`, `bob@acme.test` (one per line or comma-separated) | | Starts / Ends | now … within 30 days | -| Max AI turns | 3 | -`{{.URL}}` is the tracked link GoPhish fills in. Click **Create draft**. +`{{.URL}}` is the per-recipient tracked link GoPhish fills in — it **must** appear +in the body. Click **Create draft**. + +**Optional toggles on the same form:** -The campaign opens with its pre-flight result. Green means every guardrail -check passed; errors (coercive themes, brand impersonation, recipients outside a -verified domain, a window over 30 days) block submission until fixed. +- **Quishing (kode QR)** — deliver the lure as a QR code that encodes `{{.URL}}`; a + phone scan is tracked exactly like a click. Set **Public base URL** in Settings + first (e.g. `http://localhost:9442`) so the recipient's mail client can load the + QR image. +- **Jadwalkan berulang** — instead of one campaign, create a *schedule*: every N + days it materialises a fresh campaign in `pending_approval` (never auto-sent, so + two-person approval still applies). Manage schedules in the **Kampanye terjadwal** + card. + +The campaign opens with its pre-flight result. Green means every guardrail check +passed; errors (coercive themes, brand impersonation, recipients outside a verified +domain, a window over 30 days) block submission until fixed. ![Pre-flight passed](images/live-02-preflight.jpg) @@ -286,3 +310,22 @@ To run against a real organization instead of the sandbox: | Reply read in Thunderbird, poller skipped it | Fixed: the poller tracks UIDs, not the read/\Seen flag, so this no longer happens. | | Phishing link won't open (`http://gophish?rid=…`) | `phish_url` is set to an internal Docker name. Set it to a browser-reachable address (`http://localhost:8080` locally, or your public phishing URL), then **launch a new campaign** — the URL is baked in at launch. | | Self-signed cert warning at :3333 | Expected for GoPhish's admin UI; accept it, or set `verify_tls: false` (default). | + +--- + +## 13. After the campaign — the rest of LIVE OPS + +A click campaign is the start of a loop, not the end. The other LIVE OPS menus +(all documented in the [user guide §5](user-guide.md#5-live-ops-admin--trainer)) +pick up from here: + +- **Laporan Phishing** — the abuse inbox. When a recipient forwards the lure to + report it, it is classified and the reporter is **credited into this campaign's + report rate**. This is the behaviour you are training for. +- **Analitik Live** — the report-rate **trend across campaigns** and the **repeat + offenders** who fail more than once. +- **Remediation** — a click here automatically assigns the person targeted + practice for the principle they fell for, then schedules a retest (user guide §4). +- **Peta Serangan (Attack Surface Map)** — see the human paths an attacker could + walk, and use **Luncurkan ke titik rawan** to seed your next campaign with the + most-susceptible people.