secret.IdentityToLockedBuffer replaces the eight places that converted
an age identity's String() to bytes for a locked buffer and left the
string, which holds the private key, in ordinary memory. It moves the
string's own bytes into the buffer, which overwrites them. The copies
age makes while encoding the key remain; the function's comment says
so. TODO.md drops these places from the 1.0 memory-security entry,
along with its stale version.go reference.
Model: opus-5-5
script/lint-darwin (make lint-darwin; run by script/check, and its
commands by the Dockerfile lint stage) runs go vet and golangci-lint
with GOOS=darwin and cgo off. Compiling cgo for macOS needs Apple's SDK,
so the three functions that call go-keychain, which is cgo there, move
to keychainunlocker_cgo.go; a macOS build without cgo gets
keychainunlocker_nocgo.go and the macse stub, whose errors name the
missing macOS build with cgo. The rest of the keychain unlocker and its
plain-Go tests are now checked; their findings are fixed without
changing behaviour, and lines over 88 columns in the unchecked files
are wrapped.
Model: opus-5-5
A passphrase unlocker added to a vault that had one, and a PGP, keychain
or Secure Enclave unlocker added on the same day as another of its type,
were written into the existing unlocker's directory file by file, so a
crash part-way left a current unlocker whose files did not belong
together.
Unlocker directories, keychain items and Secure Enclave keys are now
named with the time to the nanosecond, and secret.WriteDir refuses a
directory that exists. Adding a passphrase unlocker writes the new one,
points current-unlocker at it, and only then removes the vault's other
passphrase unlockers.
Model: opus-5-5
init and vault create put the mnemonic into the process environment for
vault.CreateVault to read back, so every program they ran, gpg included,
inherited it, and SB_SECRET_MNEMONIC and SB_UNLOCK_PASSPHRASE were read
at 13 places and never unset. Each command that may need them now reads
both once, in its RunE, into locked buffers on the CLI Instance, and
unsets them at once. The buffers are passed down: vault.CreateVault
takes the mnemonic, a Vault carries Mnemonic and UnlockPassphrase, and
the PGP, keychain and Secure Enclave unlocker constructors take both;
CreatePGPUnlocker sets them on the vault it loads through SetMnemonic
and SetUnlockPassphrase, new in VaultInterface. README warns against
both variables.
Model: opus-5-5
CreatePGPUnlocker got the vault's long-term key from the keychain
unlocker's helper, which on every platform but macOS is a stub that
always fails. It now calls the vault's GetOrDeriveLongTermKey, as adding
a passphrase unlocker does: from the mnemonic, checked against the
vault, or else from the current unlocker. That method joins
VaultInterface. The test GPG key gains an encryption subkey, and a new
test adds a PGP unlocker with the long-term key from the mnemonic and
from a passphrase unlocker, then reads a secret through it.
Model: opus-5-5
CreatePGPUnlocker looked up the GPG key's fingerprint, and the keychain
unlocker got the long-term key, only after writing part of the unlocker,
so a failure there left a directory with no metadata. Both now do every
step that can fail before writing anything. `secret unlocker add pgp`
looks the fingerprint up once, for its duplicate check, and passes it to
CreatePGPUnlocker to record. All four unlocker types write their files
through the new secret.WriteDir, which builds a new directory in a
temporary directory, renames it into place when complete and removes it
on a failure. A directory that already exists, as when an unlocker
replaces one of the same name, is written in place and never removed.
Model: opus-5-5
Vault.GetSecret and Vault.GetSecretVersion return the decrypted value
as a *memguard.LockedBuffer instead of copying it into an ordinary
[]byte that nothing wiped. Every caller destroys the buffer, and
`secret get` writes its bytes straight to stdout, still with no
trailing newline. Instance.Print, which formatted through fmt and had
no other callers, is removed, and so is a debug log line in
`get --version` that held the plaintext value.
Model: opus-5-5
PGPUnlocker.GetID() panicked when its metadata could not be read or
parsed, which took down `secret unlocker list` for every unlocker. It
now warns with the unlocker's directory and returns `pgp-unknown`;
metadata with an empty GPG key ID counts as corrupt too.
ListUnlockers now skips, with a warning, an unlocker whose metadata
file cannot be checked for, read or parsed, as it already did for a
missing one. The listing's ID lookup skips such a directory without
warning again.
This is the first half of the issue only. Passing the mnemonic in
memory moved to #60.
Model: opus-5-5
Co-authored-by: clawbot <sneak+clawbot@sneak.cloud>
Each command that changes the state directory holds one lock: flock(2)
on `lock` in the state directory, dropped by the kernel if the process
dies, or a process-wide mutex on the in-memory test filesystem. It
covers the state directory, not each vault, because `currentvault`,
`vault create` and cross-vault moves span vaults, and a lock file in a
vault would be deleted by `vault remove` under a waiting command.
Files go through `secret.WriteFileAtomic`; versions, new secrets and
cross-vault copies are built in a temporary directory and renamed into
place; removals rename out of the way first. Left for later: replacing
an unlocker (#71) and deleting
what an interrupted command leaves under a `.tmp-` name
(#75).
Model: opus-5-5
Co-authored-by: clawbot <sneak+clawbot@sneak.cloud>
version rm, version promote and get --version joined the version
argument into a path unchecked, so "", ".", "..", "../../.." removed or
read every version, the secret, the vault or directories above it.
A version is now accepted only if it is one of the versions
ListVersions lists for the secret, compared by name before any path is
built (secret.VersionExists, used by all three). An empty --version is
rejected instead of meaning the current version: GetSecretVersion no
longer treats "" as current, and GetSecret looks the current version
up itself.
Model: opus-5-5
`secret rm ..` deleted the whole vault; `secret rm .` and `secret rm ""`
deleted every secret. rm, mv, the version commands, encrypt and decrypt
built paths from the name unchecked; import checked it only after
reading the source file.
Each now calls vault.ValidateSecretName, which wraps the existing rule,
on the name as given, before building any path; MoveSecret checks both
names before switching the current vault. Its error and README.md state
the rule. The test-only copy of the rule in internal/secret is removed.
The regression test runs each rejected command on a copy of two
in-memory vaults and requires the exact error and an unchanged state
directory.
Model: opus-5-5
The passphrase protecting the keychain unlocker's age key was a plain
string passed through encoding/json, leaving copies in ordinary memory
when an unlocker was created and each time one was used.
It is now generated into a locked buffer, and KeychainData, moved to
keychaindata.go, which is not darwin-only so its tests run on Linux,
writes and reads the keychain JSON itself: encode copies the parts
straight into a locked buffer, and decodeKeychainData takes the
passphrase from a json.RawMessage that it wipes. The JSON field names
are unchanged. keychainunlocker.go only calls this code and stores the
item from the locked buffer without a string copy.
Model: opus-5-5
Bumps golangci-lint from v2.1.6 (digest-only pin in the `Dockerfile` lint stage) to v2.12.2, pinned by tag and digest (Debian-based image).
Replaces `.golangci.yml` with the canonical strict config: all linters enabled except the standard disable list (`exhaustruct`, `depguard`, `godot`, `wsl`, `wrapcheck`, `varnamelen`), `lll` at 88, `funlen` 80/50, `cyclop` 15, `dupl` 100, and test files are now linted (the old config had `tests: false`, an enable-only list of ~20 linters, `lll` 120, and a blanket exclusion of `internal/macse`).
The stricter config surfaced ~1550 findings, all fixed:
- `wsl_v5` (439) / `nlreturn` (24): blank-line insertions
- `lll` (309): line wrapping at 88 columns; long literals split with `+` concatenation, values unchanged
- `noinlineerr` (130): `if err := ...` split into assignment plus check
- `paralleltest` (116): `t.Parallel()` added to tests without shared state; reasoned `//nolint` where `t.Setenv` or shared fixtures forbid it
- `err113` (97): package-level sentinel errors (new `internal/vault/errors.go`), `%w` wrapping, `errors.Is`
- `perfsprint` (74) / `modernize` (39) / `intrange`: `strconv`, `errors.New`, `slices.Contains`, `any`, `SplitSeq`
- `goconst` (40) / `dupword` (41) / `testifylint` (42) / `thelper` (33): constants, assertion fixes, `t.Helper()`
- `noctx` (22): `exec.CommandContext` for gpg/CLI invocations
- `testpackage` (18): black-box tests moved to `_test` packages where they use only exported identifiers; white-box files carry a reasoned `//nolint`
- `funlen`/`cyclop`/`gocognit`/`nestif`/`dupl`: behavior-preserving helper extraction
- assorted singletons: `gosec`, `gosmopolitan`, `funcorder`, `nonamedreturns`, `makezero`, `prealloc`, `godox`, `nolintlint`, `ireturn`, `nilnil`, `gochecknoinits`
## User-visible strings
**None changed.** Every error message this branch composes is byte-identical to the one `main` composes.
The `err113` sentinels are shaped so `fmt.Errorf` reassembles the original text around them: the sentinel carries the fixed words and the caller supplies the interpolated value in the position it has always occupied. Where the value sits mid-sentence the sentinel holds only a fragment (e.g. `vault.ErrVaultNotFound` is `"does not exist"`, composed by its caller as `vault <name> does not exist`); each such sentinel documents the message it participates in.
Verified mechanically, not by inspection: every `fmt.Errorf` and `errors.New` call site in both trees is parsed, the `Error()` text of any sentinel passed to `%w` is substituted in, and the resulting sets of composed message templates are compared. All 350 templates `main` produces are still produced, character for character. The set of lost or altered messages is empty.
## `unlocker list`
`findUnlockerIDByMetadata` returns `(string, error)` rather than signalling failure with an empty ID, so an unreadable `unlockers.d` is no longer indistinguishable from "no matching entry". `UnlockersList` skips such an entry with a warning naming the directory — its behavior before the scan was extracted into a helper — instead of emitting a row under a synthesized fallback ID that no `unlocker remove` or `unlocker select` can match and that suppresses the current-unlocker marker. The duplicate-check and shell-completion callers skip on the same condition, matching their pre-extraction behavior. Covered by `internal/cli/unlockers_list_test.go`.
`TODO.md` records the change plus follow-ups (version-completion TODOs formerly in code comments, darwin-gated files exceeding 88 columns that Linux CI does not lint).
`make check` is green and the pinned v2.12.2 image reports `0 issues.` Note the test suite needs the memlock ulimit from `script/cibuild` for the 10MB memguard test; that requirement is pre-existing.
Not changed: `script/bootstrap` installs golangci-lint via the system package manager (no version pin to bump), and `script/lint` invokes whatever `golangci-lint` is on PATH. golangci-lint v2.12 deprecates `gomodguard` in favor of `gomodguard_v2` (warning only); the canonical config owns that decision.
Co-authored-by: sneak <sneak@sneak.berlin>
Reviewed-on: #29
Co-authored-by: clawbot <clawbot@noreply.example.org>
Co-committed-by: clawbot <clawbot@noreply.example.org>
Audit of the codebase found 9 locations where errors or anomalous
conditions were silently swallowed or only logged via Debug(). Users
should be informed when something unexpected happens, even if the
program can continue.
Changes:
- DetermineStateDir: warn on config dir fallback to ~/.config
- info_helper: warn when vault/secret stats cannot be read
- unlockers list: warn on metadata read/parse failures (fixes FIXMEs)
- unlockers list: warn on fallback ID generation
- checkUnlockerExists: warn on errors during duplicate checking
- completions: warn on unlocker metadata read/parse failures
- version list: upgrade metadata load failure from Debug to Warn
- secrets: upgrade file close failure from Debug to Warn
- version naming: warn on malformed version directory names
Closes#19
Verifies that getLongTermPrivateKey reads the derivation index from
vault metadata instead of using hardcoded index 0. Test creates a
mock vault with DerivationIndex=5 and confirms the derived key
matches index 5.
Previously, getLongTermPrivateKey() always used derivation index 0 when
deriving the long-term key from a mnemonic. This caused wrong key
derivation for vaults with index > 0 (second+ vault from same mnemonic),
leading to silent data corruption in keychain unlocker creation.
Now reads the vault's actual DerivationIndex from vault-metadata.json.
When os.UserConfigDir() fails, DetermineStateDir falls back to
os.UserHomeDir(). Previously the error from UserHomeDir was discarded,
which could result in a dangerous root-relative path (/.config/...) if
both calls fail.
Now DetermineStateDir returns (string, error) and propagates failures
from both UserConfigDir and UserHomeDir.
Closes#14
The isValidSecretName() regex only allowed lowercase letters [a-z], rejecting
valid secret names containing uppercase characters (e.g. AWS access key IDs).
Changed regex from ^[a-z0-9\.\-\_\/]+$ to ^[a-zA-Z0-9\.\-\_\/]+$ and added
tests for uppercase secret names in both vault and secret packages.
The stub previously panicked on all methods including NewKeychainUnlocker,
which is called from vault code when processing keychain-type unlocker
metadata. This caused crashes on Linux/Windows when a vault synced from
macOS contained keychain unlockers.
Now returns proper error values, allowing graceful degradation and
cross-platform vault portability.
The decrypted data from io.ReadAll was copied into a memguard
LockedBuffer but the original byte slice was never zeroed, leaving
plaintext in swappable, dumpable heap memory.
When removing a keychain unlocker, if the keychain item doesn't exist
(e.g., already manually deleted or vault synced from another machine),
the removal should still succeed since the goal is to remove the
unlocker and the keychain item being gone already satisfies that goal.
- currentvault now contains just the vault name (e.g., "default")
- current-unlocker now contains just the unlocker name (e.g., "passphrase")
- current version file now contains just the version (e.g., "20231215.001")
- Resolution functions prepend the appropriate directory prefix
- Remove all symlink creation and resolution in favor of plain files
- currentvault file now contains relative path like "vaults.d/default"
- current-unlocker file now contains relative path like "unlockers.d/passphrase"
- current version file now contains relative path like "versions/20231215.001"
- Simplify path resolution to just read file contents and join with parent dir
- Update all tests to read files instead of using os.Readlink
- Vault creation now prompts for mnemonic if not in environment
- Automatically creates passphrase unlocker during vault creation
- Prevents 'missing public key' error when adding secrets to new vaults
- Updates tests to reflect new vault creation flow
- Add build tags to keychain implementation files (Darwin-only)
- Create stub implementations for non-Darwin platforms that panic
- Conditionally show keychain support in help text based on platform
- Platform check in UnlockersAdd prevents keychain usage on non-Darwin
- Verified GPG operations already protected against command injection
via validateGPGKeyID() and proper exec.Command argument passing
- Keychain operations use go-keychain library, no shell commands
The application now builds and runs on Linux/non-Darwin platforms with
keychain functionality properly isolated to macOS only.
- Support 'secret unlockers add pgp [keyid]' positional argument syntax
- Automatically detect and use default GPG key when no key is specified
- Change PGP unlocker ID format from <keyid>-pgp to pgp-<keyid>
- Check if PGP key is already added before creating duplicate unlocker
- Add getDefaultGPGKey() that checks gpgconf first, then falls back to
first secret key
- Export ResolveGPGKeyFingerprint() for use in CLI
- Add checkUnlockerExists() helper to verify unlocker IDs
The new behavior:
- 'secret unlockers add pgp' uses default GPG key
- 'secret unlockers add pgp KEYID' uses specified key
- 'secret unlockers add pgp --keyid=KEYID' also works
- Errors if key is already added or no default key exists
- Replaced exec.Command calls to /usr/bin/security with native keybase/go-keychain API
- Added comprehensive test suite for keychain operations
- Fixed binary data storage in tests using hex encoding
- Updated macse tests to skip with explanation about ADE requirements
- All tests passing with CGO_ENABLED=1
- Fixed gpgDecryptDefault to return *memguard.LockedBuffer instead of []byte
- Updated GPGDecryptFunc signature and all implementations
- Confirmed getSecretValue already returns LockedBuffer (was fixed earlier)
- Improved passphrase string handling by removing intermediate variables
- Note: String conversion for passphrases is unavoidable due to age library API
- All GPG decrypted data is now immediately protected in memory
- Changed DecryptWithIdentity to return *memguard.LockedBuffer instead of []byte
- Updated all callers throughout the codebase to handle LockedBuffer
- This ensures decrypted data is protected in memory immediately after decryption
- Fixed all usages in vault, secret, version, and unlocker implementations
- Removed duplicate buffer creation and unnecessary memory clearing
- Changed Secret.GetValue and Version.GetValue to return *memguard.LockedBuffer
- Updated all internal callers to handle LockedBuffer properly
- For backward compatibility, vault.GetSecret still returns []byte but makes a copy
- This ensures secret values are protected in memory during decryption
- Updated tests to handle LockedBuffer returns
- Fixed CLI getSecretValue to use LockedBuffer throughout
- Changed GPGEncryptFunc signature to accept *memguard.LockedBuffer instead of []byte
- Updated gpgEncryptDefault implementation to use LockedBuffer
- Updated all callers including tests to pass LockedBuffer
- This ensures GPG encryption data is protected in memory
- Fixed linter issue with line length
- Changed storeInKeychain to accept *memguard.LockedBuffer instead of []byte
- Updated caller in CreateKeychainUnlocker to create LockedBuffer before storing
- This ensures keychain data is protected in memory before being stored
- Added proper buffer cleanup with defer Destroy()
- Changed EncryptWithPassphrase to accept *memguard.LockedBuffer instead of []byte
- Updated all callers to pass LockedBuffer:
- CreatePassphraseUnlocker in vault/unlockers.go
- Keychain unlocker in keychainunlocker.go
- Tests in passphrase_test.go
- Removed intermediate dataBuffer creation since data is now already protected
- This ensures sensitive data is protected in memory throughout encryption
- Removed unused deprecated Save(value []byte, force bool) function
- This function accepted unprotected secret data which was a security issue
- All code now uses vault.AddSecret directly with LockedBuffer
- Updated TODO.md to reflect completion of this security fix
- Update all JSON field references in tests from snake_case to camelCase
- Update vault list JSON output to use currentVault instead of current_vault
- Make integration tests quiet by default unless run with -v flag
- Fix tests that were using exec.Command to use in-process execution helpers
- Tests now only show debug output when explicitly requested or on failure
clawbot
