---
{
  "title": "Digital signatures",
  "summary": "Learn what digital signatures verify, why trusted public keys matter, and what a valid signature cannot prove about software safety.",
  "topic": "foundations",
  "level": "Intermediate",
  "updated": "2026-10-12",
  "units": [
    9
  ],
  "aliases": [],
  "sources": [
    {
      "title": "NIST: Cryptographic Standards and Guidelines",
      "url": "https://csrc.nist.gov/projects/cryptographic-standards-and-guidelines"
    },
    {
      "title": "NIST: Digital Signature Standard, FIPS 186-5",
      "url": "https://csrc.nist.gov/pubs/fips/186-5/final"
    }
  ],
  "published": "2026-10-02",
  "editorialVersion": 2
}
---

## Verification links content to a signing key

A digital signature supports verification that particular content was signed using private key material corresponding to a public key, under the signature scheme's assumptions. Modifying the signed content should cause verification to fail. Associating that key with a real organization or person requires an additional trust process.

A signature does not hide the message. Anyone with access to the content may still read it unless a separate confidentiality mechanism protects it. [[asymmetric-encryption]] explains the wider public-key setting, and [[hashing]] explains a related building block used in signature schemes.

## A software updater needs more than a green tick

Imagine a fictional academy distributes a desktop learning tool. Its updater receives a package and a signature. The updater must verify the signature using an appropriate trusted publisher key and apply the protocol's validation rules before accepting the release.

If the package arrives with an unfamiliar public key and the updater simply trusts that key, an attacker could supply both. The signature may verify mathematically while the package has no established relationship to the academy. “Signature valid” needs context about which key was trusted and why.

Now imagine an old, correctly signed release contains a known security flaw. Its signature does not turn it into a safe current release. Version and rollback policy address another part of update security. Authentication of content and assessment of its suitability are different decisions.

## A valid signature is not a safety review

Signed software can contain defects or malicious behavior. A compromised publisher may sign harmful content, or an attacker may misuse the publisher's signing capability. Verification tells you something precise about content and key relationships; it is not a guarantee of quality, intent, or absence of vulnerabilities.

For the fictional updater, this means signature verification belongs alongside controlled release approval, testing, version policy, and monitoring. A security review should ask how the package became authorized for release rather than ending when a cryptographic check succeeds.

## Key custody affects the evidence

Suppose the academy discovers that its signing account was used unexpectedly. Responders need to determine which releases were signed, whether the underlying key or signing service was compromised, and how relying devices will receive updated trust information. Simply generating a new key does not explain what devices should do with previously accepted releases.

[[key-management]] addresses the lifecycle and permissions involved. [[digital-certificates]] explains a common way of binding identities and public keys, with validation requirements of its own.

Claims about accountability also depend on custody and process. If many people share unrestricted access to a signing credential, a signature alone cannot distinguish their individual actions. Describe the evidence and its limits rather than making universal legal claims about non-repudiation.

## Practice reading the conclusion precisely

Use a release in a designated training environment with known test keys. Verify the original content, change a byte, and observe the verification result. Then try the original content with a different public key.

Write a sentence for each observation. A careful answer states whether verification succeeded under the chosen key and scheme, which trust assumptions were supplied, and which questions about the program remain unanswered. That precision is more useful than labeling every signed file trustworthy.