One delivery manifest, two deliberately separate verification paths: a content
identifier addresses content; a raw-file digest checks the delivered bytes. Verification states
are shown exactly as participants will meet them — matched, mismatched, absent, stale.
Design candidate. Illustrates the planned receipts stage's release
criteria with synthetic content. Not a live API or storage commitment — see the
vision index for scope.
Journey view · one delivery manifest
The manifest card
Everything the receipts stage requires from a delivery manifest, readable in one
view. Submission time and established creation time render as two distinct facts; retention and
access conditions are part of the receipt, not footnotes.
2026-09-14 09:20 UTC — independently established creation time
Artifact location
ipfs://bafy…synthetic · retained by @kofi and @aria until acceptance + 90 days
Media type · size
application/zstd · 4.1 KiB
Evidence references
Two verification runs, each with its own manifest
Access conditions
Public read · no credentials required · nothing private included
Criteria addressed
Manifest fields: project and milestone identifiers, scope version, delivery revision, author
attribution, artifact locations, content identifiers, declared media types, sizes, integrity
algorithm and digest, evidence references, publication/access conditions · submission timestamp
distinguished from established creation time · retention agreed and named.
Journey view · two verification paths
CID and digest answer different questions
A CID includes information about the represented content and its encoding — it is
not always the ordinary checksum of the original file. The UI keeps the two paths apart and names
what each one actually establishes.
Content addressCID
Representation
CIDv1, raw codec — names content and encoding
Value
bafk…synthetic-cid
Establishes
Which bytes this manifest refers to, addressable by content
Does not establish
Quality, authorship, rights or acceptance
Criteria addressed
Distinguish a CID from a raw-file digest · content addressing supports integrity checking, not
acceptance.
Byte checkraw digest
Algorithm
SHA-256 over the retrieved bytes
Expected
9f2c…synthetic
How
Independent verification instructions, executable by anyone with retrieval access
Establishes
That retrieved bytes match this manifest's digest
Criteria addressed
Integrity algorithm and digest declared · validate artifact retrieval without letting supplied
URLs become unrestricted server requests.
Journey view · verification states
Every state reads as evidence, never as decoration
Color is paired with words; each row says what was compared and what that means.
A stale revision never renders like a success, and an absent provider is a different state from
a hash mismatch.
matchedrev 2
Retrieved bytes hash to the expected digest. Compared: 4.1 KiB retrieved
2026-09-15 14:05 UTC → 9f2c…synthetic.
mismatchrev 1
Retrieved bytes differ from the declared digest. This is a delivery defect,
not an acceptance decision.
absentrev 2
No provider served the content at check time. A CID alone does not provide
storage; retention is a named commitment.
stale revisionrev 1
Bytes match — but rev 2 exists and supersedes this delivery. Matching bytes
of a stale revision are not a current delivery.
Criteria addressed
Release criteria: detect mismatched bytes, unsafe references, absent content and stale revisions ·
preserve the agreed access policy · test missing providers and restore paths.