fix(controlplane): return error on short attestation state ciphertext

[beiryu]

Fixes #3600

decrypt in app/controlplane/pkg/biz/attestationstate.go checked the salt and IV lengths but not the length of the remaining data before comparing it with the magic prefix. A stored value of 32 to 44 bytes passed the earlier checks and then caused a slice-out-of-range panic.

This change adds a len(ciphertext) < len(magic) check before the comparison, so a truncated or corrupted value now returns an error.

This does not cover the second point in the issue (moving to an authenticated cipher such as AES-GCM). That needs a format change and a migration path for existing rows, so it is left for a separate change.

Testing

• Added TestDecryptShortInput in attestationstate_test.go. It calls decrypt with inputs of salt + IV only, one byte after the IV, and one byte short of the magic prefix. It checks that there is no panic and that an error is returned.
• Added TestEncryptDecryptRoundTrip to check that normal encrypt and decrypt still work.
• The new tests and the project checks pass.

The decrypt function could panic on a truncated database value. I wanted it to return an error, so one bad row does not crash the request.

[chainloop-platform]

PR validation — ✅ 3 passing

Status	Policy	Material	Messages
✅ Passed	pr-user-story-linked	pr-info	-
✅ Passed	pr-min-approvals	pr-info	-
✅ Passed	pr-description-required	pr-info	-

_{View attestation ↗}

AI Session Checks — ⏭️ bypassed by label

AI Coding Session Check Bypassed

This PR carries the skip-ai-session label, so the AI coding session check was bypassed.

Learn more about Chainloop Trace.

Security Checks — ✅ 5 passing

✅ secret-scan

Status	Policy	Messages
✅ Passed	secrets-detection	-

✅ sast-scan

Status	Policy	Messages
✅ Passed	owasp-top10-2025	-
✅ Passed	sast	-
✅ Passed	cwe-top25	-
✅ Passed	cwe-top26-40-cusp	-

Scans not applied (3)

Scan	Reason
vulnerability-scan	no manifest/lockfile changed
github-actions-scan	no workflow files changed
iac-scan	no IaC files changed

_{View attestation ↗}

Security context

✅ Nothing this change touches has a recorded security-fix history.

_{View in Chainloop ↗ · How this works ↗}

---

Powered by Chainloop and Chainloop Trace

[cubic-dev-ai]

No issues found across 2 files

_{View guided diff | Re-trigger cubic}

[migmartri]

Thanks for the contribution, could you please sign make sure the commits are signed?

[beiryu]

Thanks! The commit is now signed (SSH signature, shows as Verified). The content and the DCO sign-off are unchanged.

[migmartri]

Sorry, there seems to be some conflicts, could you resolve them please, thanks!

[beiryu]

@migmartri sorry, I had to push one more commit after your approval (b30a8ac), so the approval was dismissed.

Test (controlplane) failed in TestAttestationStateUseCase/TestRead: the "tampered with" case stores 41 bytes, which leaves 9 bytes after the salt and IV. That is shorter than the magic prefix. On main, ciphertext[:len(magic)] reads into the slice's spare capacity, so the case ended up as ErrInvalidPassphrase. My new ciphertext too short error changed that.

The commit now returns NewErrInvalidPassphrase() for the short case, so a truncated or tampered row is reported the same way as before (no panic), and TestDecryptShortInput asserts that. TestAttestationStateUseCase passes locally against Postgres. Could you take another look when you have a moment?