Evidence practice / 3 MIN READ
A matching file hash does not establish who approved the request
Compare three fictional evidence files to separate byte integrity, filenames, formatting and the approval claim a reviewer still needs to verify.
On this page
A file hash can support a comparison of captured bytes. It does not, by itself, identify the author of a statement or establish that the statement describes an authorized action.
NIST describes hash functions as producing a fixed-length digest from a message. This original fictional exercise uses SHA-256 to make the comparison visible. It is career practice, not a digital-forensics procedure, legal chain-of-custody standard or assurance opinion.
Define the bytes being compared
File A contains the UTF-8 text REQ-17 approved, with no newline. File B has a different filename but exactly the same content bytes. File C contains that text followed by one line-feed byte. No byte-order mark is included.
| File | Content length | First twelve characters of SHA-256 |
|---|---|---|
| A | 15 bytes | 235fa2783219 |
| B | 15 bytes | 235fa2783219 |
| C | 16 bytes | c96dce49e565 |
The digest prefixes are abbreviated display labels, not sufficient verification values. Each complete digest below is displayed on two lines for readability: concatenate its lines with no separator to obtain the 64-character value. The digests for A and B are identical:
235fa27832199f2e9524c88357615a34
2e4d013ebeb7a57c9c0ac5a3c4d9a2aa
C's complete digest is:
c96dce49e5651c5014ccfb166a71ed36a
0462ab9493f4a260dc1c4a429a86522
These values hash only the stated content bytes, not a filename or filesystem timestamp. A and B's full-digest match is consistent with the unchanged-content copy supplied in this case. Renaming a file does not change these content bytes. C differs because it adds a byte; that difference alone does not establish fraud or a changed approval decision.
Reproduce the small example
The Python hashlib reference documents SHA-256 and hexadecimal digest output. With Python installed, the following uses invented byte strings only:
import hashlib
inputs = (
b"REQ-17 approved",
b"REQ-17 approved\n",
)
for data in inputs:
digest = hashlib.sha256(data)
print(len(data), digest.hexdigest())
Do not normalize line endings or resave a captured original silently before comparing it. If an approved process creates a normalized derivative, identify that derivative separately and retain the relationship to the original.
Write the narrower conclusion
Try: “B's complete SHA-256 digest matches A's recorded content digest. C's digest differs; the supplied construction identifies an added line feed. The text in all three files says approved, but none identifies an approver, authority, decision time or source workflow. Those approval facts remain unverified.”
You also need a trustworthy reference. If someone can replace both a file and its supposed reference digest, matching the two does not resolve that provenance problem. Record the source, capture method, time and object covered by the digest under the applicable authorized process.
Keep the fictional files, full digests and limited note as practice. Compare file-integrity preparation resources. Use the evidence-reliability exercise for a different question: whether a report represents the population you intend to examine.
PUT IT INTO PRACTICE
GRC Career Transition Guide
Turn this guidance into a focused role brief, experience inventory and application plan.