Cards¶
issue SEQ renders a human-readable evidence card from a record. The card is derived data: it can be rebuilt
from the ledger at any time, and the ledger remembers its hash.
What a card is built from¶
- the record at
SEQ(status, sources, reproduction level, claim boundary, reasons); - the frozen protocol it names (id, version, hash);
- the optional
cardobject inside the record body, which carries the remaining evidence-card fields (operator, source name and URL, domain, claim type, access date, hashes of the sanitized report, known limitations, and so on).
Fields derived from the record always win: a card key that contradicts them is rejected with CARD_CONFLICT.
Missing required fields are rejected with MISSING_FIELD and named; any other validation failure is
CARD_INVALID. The card is checked by the evidence-card validator before anything is written.
Files¶
.protocolith/cards/PLITH-<protocol_id>-<date of the record>.json
.protocolith/cards/PLITH-<protocol_id>-<date of the record>.svg
Change the folder with --out DIR. The JSON file is the canonical bytes of the card with no trailing newline,
so the SHA-256 of the file equals card_sha256 in the issue entry. Issuing the same record again writes the
same bytes (and appends another issue entry). If the file exists with different content, the result is
CARD_EXISTS and nothing is overwritten. raw_payload_committed is always false.
Rule R9¶
Verification rebuilds the card from the record and compares its hash with card_sha256. If the record, the
protocol or the evidence-card validator's output changed since, verification reports BAD_CARD.
Reading a card¶
Read three fields first: result_status, reproduction_level and claim_boundary. A green card means one
narrow claim passed under one frozen protocol, not that a system is good. Most cards produced from a single
local run are not independently reproduced. The older card documentation is in the
archive; the full list of fields is in EVIDENCE_CARD.