1Why does Abacus ask for PGP when Tor already hides the route?
To understand why vendors and buyers use public keys instead of plaintext addresses, think of Tor as hiding the road you travel, not the contents of the package. Onion routing conceals your location and destination, but the message itself remains readable to anyone who reaches the endpoint — a market server compromise, a logged database, or a seized server image all expose plaintext directly.
Abacus asks for PGP because encryption is endpoint protection, not transport protection. When your order message travels through Tor, it is encrypted in transit between relays, but once it arrives inside the market's conversation system, it sits as ordinary text unless you personally encrypt it first. A vendor reading your order sees exactly what you typed — and whatever the server stores is what a later seizure could recover.
The key differential: who holds the decryption key
When you encrypt with a vendor's public key, you produce ciphertext that only the vendor's private key can open. The market's server never sees the plaintext, never stores it as readable text, and cannot leak it in readable form even during a seizure. That is a meaningful property gap, not a formality.
If a buyer pastes a plaintext address into an order, the address exists as clear text on the market server from that moment onward — regardless of how safely the Tor circuit was built. Transport encryption cannot undo what you wrote before it reached the relay.
Tor answers a different question: "Can someone see where I am and where I am going?" PGP answers: "If this message is captured at its destination, can anyone read it?" Needing both is not redundancy; they cover different failure points.
- Analogy: Tor is an unmarked truck
2What actually goes wrong when a buyer pastes a plaintext address into an order?
A plaintext shipping address turns an otherwise private market order into a stored confession. Tor hides the route between you and Abacus, but it does not hide the order text from the marketplace database, a seized server, or anyone who later reads that server’s contents. Once the address exists as readable text inside an order, it can be tied to your buyer account, your message history, and your payment details in a single screen.
The typical failure chain is short and practical. A buyer pastes a name, street, and postal code directly into the order notes. The market stores that text on its own infrastructure, often alongside order timestamps and account identifiers. If law enforcement later takes control of the server, that data is readable instantly. What Tor protected in transit is gone, because the destination itself kept the secret in plain sight.
Do not assume a market will delete addresses after delivery. Retention depends on server configuration and market policy, and neither is something you can verify from your side. The only safe assumption is that whatever you type may remain on that infrastructure longer than the order itself.
Public keys exist to solve exactly this problem. Instead of storing the delivery address as readable text, the buyer encrypts the address to the vendor’s public key. The encrypted block is meaningless to anyone without the vendor’s private key, including Abacus itself. Even a full database copy would reveal only ciphertext where the address used to be.
There is also a secondary risk beyond government seizure. A vendor who receives an address in plaintext can later deny the order arrived, and the buyer has no cryptographic record of what was actually sent. Even in a dispute, a staff member reading the plaintext address has no way to prove that specific account sent that specific text. Encryption ties the message to a key, not a person.
- Plaintext address
- a physical delivery name and location written without encryption, readable by anyone with database access
- Public key
- the half of a keypair used to encrypt a message so only the matching private key can decrypt it
- Ciphertext
- the scrambled output of encryption, unreadable without the correct private key
The point of a public key is not to hide the vendor. It is to keep the sensitive part readable only by the one party that actually needs it.
If a market form offers an optional "encrypt your address" field, leaving it empty means the site itself can read what you wrote. Treat every unencrypted field as a message to the server, not to the vendor alone.
3How do you tell a vendor's real public key from a copied or swapped one?
Before trusting a vendor key, compare its 40-character fingerprint against any past copy you saved — not just the 8-character short ID, which is easy to fake. A swapped key usually looks almost identical in a profile but produces a completely different fingerprint when you import it into your own keychain.
Why a badge is not proof
A “verified” or pinned key on a profile means only that the profile owner once placed that key in the field. It does not prove the vendor and the key belong to the same person over months, and it does not protect you from a marketplace compromise, a fake mirror, or a vendor account takeover.
The one-minute check before encrypting
- Copy the fingerprint from your saved key store Compare it character by character with the fingerprint shown on the profile and in any order history
- Look at key creation date and user IDsA genuine vendor key rarely changes from week to week; fresh keys with no past use are a red flag
- Check the profile name against the key's user IDMismatch between the handle and the key ID text often signals a copied key from another account
- Encrypt a short test line firstIf your own tool rejects the key or the fingerprint changed after import, stop and do not send an address
Fingerprint match
Compare the full 40 hex characters, not the short ID
Profile consistency
Same key on the same handle for weeks or months, not a fresh replacement
Key creation date
A vendor key created yesterday while the account claims years of sales
Past order history
If you used this vendor before, your keychain should already hold the same key without import errors
If a vendor asks you to “re-import” a new key just before payment, treat the request as a compromise signal until you confirm the old fingerprint through a separate channel.
Save every trusted vendor fingerprint outside the market — a plain text note, an offline file, or a password manager. The profile field itself can change at any moment.
FAQ
Why does Abacus ask for PGP when Tor already hides the route?
Tor hides the network path between you and the market, but it does not encrypt the content of a message or address after it reaches the server. PGP protects that stored content, so a seized or compromised server cannot read payment details even if the route was hidden.
What actually goes wrong when a buyer pastes a plaintext address into an order?
The address becomes readable to anyone who later gains access to the order database, including an attacker or a market takeover. That means funds can be traced or diverted before the buyer notices anything unusual.
How do you tell a vendor's real public key from a copied or swapped one?
You compare the key fingerprint through a separate trusted channel, not just the key text shown on one page. If the fingerprint you saved or verified earlier does not match the one on the order page, stop and do not send anything.
Why does encrypting to yourself matter even before you send anything?
It lets you keep a copy of your own encrypted message that only your private key can open, so you have proof of what you actually wrote. Without it, you lose the ability to show a vendor the exact encrypted text you sent if a dispute arises.
What should you do if encryption still feels unreliable or a key check fails?
Do not place the order until the key matches a fingerprint you trust and your encryption tool gives a clear success result. Slow down, re-verify the key from an independent source, and practice with a test message before risking any real address or payment detail.
4Why does encrypting to yourself matter even before you send anything?
Self-encryption closes the one gap that public-key messaging cannot cover by itself: proving later what you actually sent, not just that someone else could read it. If a dispute opens, the market operator or arbiter usually cannot decrypt your outbound copy of an order message — they only see ciphertext. What you can show is the encrypted copy you saved with your own public key, because only your private key can open it. Without that step, the only readable evidence is the plaintext draft sitting on your machine, which is exactly the thing you were trying not to leave behind.
Think of it as keeping a sealed duplicate inside a safe that only you control. The vendor’s public key protects the message going out; your public key protects the record coming back. When both copies exist, you can decrypt your own archive later and show the address, quantity, or note exactly as it was written — without exposing the plaintext before you need it.
What self-encryption actually gives you
- A recoverable order snapshot that does not depend on the market staying online
- A way to verify your own key still works before relying on it for anything serious
- Evidence you can decrypt locally if a vendor claims the address was different
Before encrypting anything sensitive on Abacus, encrypt a short test note to yourself and decrypt it again. If that round trip takes more than two minutes or fails, fix it before touching a real address.
Do not confuse “encrypted to your own key” with “typed into a message and marked as draft.” A draft is still plaintext unless the application explicitly encrypts it, and many do not.
Some PGP tools allow encrypting to multiple recipients simultaneously — vendor and self. Where supported, that is the cleanest approach; where not, encrypt twice in two steps.
5What should you do if encryption still feels unreliable or a key check fails?
Treat an encryption problem as a narrow diagnostic, not a personal failure or a reason to abandon PGP. The most common cause is a mismatched key pair — the public key you imported does not match the private key your client is using. In that case the message encrypts, but only the key’s real owner can decrypt it, and you may never notice until a vendor tells you the text is unreadable. Check this first: open your key manager and confirm the key ID next to the vendor’s imported key is the same one your client selected when signing or encrypting.
A full key check fail — where the fingerprint on the market page does not match the vendor’s advertised key — is a stop signal, not a request to “try harder.” Treat it like a broken seal on a package: the content may be fine, but the channel is compromised. Do not encrypt an address to that key and do not complete the order until the mismatch is explained through a channel you initiated yourself.
If a vendor publishes a new key and claims it replaces an old one, verify that change outside the market’s messaging system. A swapped key is one of the quietest ways an order can be redirected.
Quick checks before you assume the worst
- Check your keyring first Import the vendor’s key again from the profile and compare the 16-character key ID with the one already stored. A changed ID means a changed key.
- Test-encrypt to yourself Create a short message, encrypt it to your own public key, then decrypt it. If this fails, the problem is local, not the vendor.
- Confirm the fingerprint away from the market Copy the fingerprint into a note, close the market tab, and compare it with a cached copy or a vendor quote you saved earlier.
- Treat a “lost private key” as a reset You cannot recover a private key you no longer have. Generate a new pair, delete the old public key where possible, and notify any vendor you already contacted.
The urge to keep using a half-working setup because “it worked before” is the main risk at this stage. A private key that opens messages only sometimes, or a client that silently falls back to plaintext, is more dangerous than no encryption at all because it creates false confidence. If your tool cannot consistently sign, encrypt, and verify — switch tools, not standards.
Most “PGP is unreliable” problems are actually clipboard, keyring, or client-selection errors. The math is rarely the part that fails.