Nosey Parker is a command-line program that finds secrets and sensitive information in textual data and Git history.
APACHE-2.0 License
Bot releases are visible (Hide)
The scan
and github repos list
commands offer a new --github-repo-type={all,source,fork}
option to select a subset of repositories (#204).
A category mechanism is now provided for rules (#208). Each rule can have zero or more freeform text categories assigned to it. The existing rules have been updated with category information with the following meanings:
secret
: the rule detects things that are in fact secretsidentifier
: the rule detects things that are not secrets but could be used to enumerate additional resources (e.g., S3 bucket names)hashed
: the rule detects hashed payloads (e.g., bcrypt hashes)test
: the rule detects test deployment-specific payloads (e.g., stripe test keys)api
: the rule detects payloads used for API accessgeneric
: the rule is a "generic" one rather than one that detects a specific type of payload (e.g., username/password pairs)fuzzy
: the rule pattern requires matching of non-payload surrounding contextThe category information is included in output in the rules list
command.
The scan
and github repos list
commands now only consider non-forked repositories by default (#204). This behavior can be reverted to the previous behavior using the --github-repo-type=all
option.
The Alpine-based Docker image has been updated to use the alpine:latest
base image instead of alpine:3.18
(#201).
The "Blynk Organization" rules have been refined (#208). The two "Blynk Organization Client ID" and two "Blynk Organization Client Secret" variations have been subsumed by two new Blynk Organization Client Credential
rules. These new rules combine the client ID and client secret into single findings instead of reporting them as two separate findings as previous.
Several rules have been renamed (#208):
AWS S3 Bucket (subdomain style)
-> AWS S3 Bucket
AWS S3 Bucket (path style)
-> AWS S3 Bucket
Blynk Organization Access Token (URL first)
-> Blynk Organization Access Token
.Blynk Organization Access Token (URL last)
-> Blynk Organization Access Token
.Generic Password (double quoted)
-> Generic Password
Generic Password (single quoted)
-> Generic Password
Generic Username and Password (quoted)
-> Generic Username and Password
Generic Username and Password (unquoted)
-> Generic Username and Password
Google Cloud Storage Bucket (path style)
-> Google Cloud Storage Bucket
Google Cloud Storage Bucket (subdomain style)
-> Google Cloud Storage Bucket
Google OAuth Client Secret (prefixed)
-> Google OAuth Client Secret
New Relic License Key (non-suffixed)
-> New Relic License Key
particle.io Access Token (URL first)
-> particle.io Access Token
particle.io Access Token (URL last)
-> particle.io Access Token
Note that although several rules share the same name now, they all still have distinct IDs.
The default set of patterns for the existing gitignore-style path-based exclusion mechanism (scan --ignore=GITIGNORE_FILE
) has been expanded (#209). The new patterns cover test files from things like vendored Python, Node.js, and Go packages.
The gitignore-style path-based exclusion patterns (scan --ignore=GITIGNORE_FILE
) now also apply to content found within Git history, and not just paths on the filesystem (#209). When a blob is found in Git history with at least 1 associated pathname, if all of the associated pathnames match the ignore rules, the blob is not scanned.
The Rust version required to build has been bumped from 1.76 to 1.77. This is necessary to support C-string literals in the rusqlite
crate.
Published by bradlarsen 3 months ago
A prebuilt multiplatform Docker image for this release is available for x86_64 and ARM64 architectures:
docker pull ghcr.io/praetorian-inc/noseyparker:v0.18.1
Additionally, a prebuilt Alpine-based image is also available for x86_64 and ARM64 architectures:
docker pull ghcr.io/praetorian-inc/noseyparker-alpine:v0.18.1
Nosey Parker no longer crashes upon startup when running in environments with less than 4 GiB of RAM (#202).
The Base64-PEM-Encoded Private Key
rule has been refined to reduce false positives and avoid a rare performance pitfall.
Published by bradlarsen 4 months ago
A prebuilt multiplatform Docker image for this release is available for x86_64 and ARM64 architectures:
docker pull ghcr.io/praetorian-inc/noseyparker:v0.18.0
Additionally, a prebuilt Alpine-based image is also available for x86_64 and ARM64 architectures:
docker pull ghcr.io/praetorian-inc/noseyparker-alpine:v0.18.0
The README now includes several animated GIFs that demonstrate simple example use cases (#154).
The report
command now offers a new --finding-status=STATUS
filtering option (#162). This option causes findings with an assigned status that does not match STATUS
to be suppressed from the report.
The report
command now offers a new --min-score=SCORE
filtering option (#184). This option causes findings that have a mean score less than SCORE
to be suppressed from the report. This option is set by default with a value of 0.05.
A new datastore export
command has been added (#166). This command exports the essential content from a Nosey Parker datastore as a .tgz file that can be extracted wherever it is needed.
New experimental annotations export
and annotations import
commands have been added (#171). These commands allow annotations (finding comments, match comments, and match statuses) to be converted between JSON and datastore representations.
New rules have been added:
Prebuilt releases now included separate debug symbols (.dSYM or .dwp files) (#191). Having the debug symbols available makes stack traces more useful in the rare event of a crash. The Alpine-based Docker image does not include these debug symbols, as its point of existing is to provide a small distribution.
The summarize
command now includes additional columns for the assigned finding status (#196).
The vendored copy of Boost included in the internal vectorscan-sys
crate has been removed in favor of using the system-provided Boost (#150 from @seqre). This change is only relevant to building Nosey Parker from source.
The vendored copy of the Vectorscan regular expression library included in the internal vectorscan-sys
crate has been removed (#151 from @seqre). Instead, a copy of the Vectorscan 5.4.11 source tarball is included in this repository, and is extracted and patched during the build phase.
SARIF reporting format is now listed as experimental.
In the scan
and rules
command, the command-line option to load additional rules and rulesets from files has been renamed from --rules
to --rules-path
. The old --rules
option is still supported as an alias, but this is deprecated and will be removed in the v0.19 release.
The rules list
command now includes additional fields when using JSON format (#161).
The vectorscan
and vectorscan-sys
crates have been split off into a separate project with crates published on crates.io (#168).
The scan
command is now more conservative in its default degree of parallelism (#174). Previously the default value was determined only by the number of available vCPUs. Now the default value is additionally limited to ensure at least 4 GiB of system RAM per job.
The scan
command now records its results incrementally to the datastore instead of in one enormous transaction (#189). Now, results are recorded in transactions about every second. This helps avoid complete loss of scan results in the rare event of a crash.
A rare crash when parsing malformed Git commit timestamps has been fixed by updating the gix-date
dependency (#185).
Upon noseyparker
startup, if resource limits cannot be adjusted, instead of crashing, a warning is printed and the process attempts to continue (#170).
The prepackaged releases and binaries produced by the default settings of cargo build
should now be more portable across microarchitectures (#175). Previously, the builds would be tied to the microarchitecture of the build system; this would sometimes result in binaries that were not portable across machines, particularly on x86_64.
The --ignore-certs
command-line option is now a global option and can be specified anywhere on the command line.
Published by bradlarsen 8 months ago
A prebuilt multiplatform Docker image for this release is available for x86_64 and ARM64 architectures:
docker pull ghcr.io/praetorian-inc/noseyparker:v0.17.0
Additionally, a prebuilt Alpine-based image is also available for x86_64 and ARM64 architectures:
docker pull ghcr.io/praetorian-inc/noseyparker-alpine:v0.17.0
A new --ignore-certs
command-line option has been added to the scan
and github
commands. This option causes TLS certificate validation to be skipped (#125; thank you @seqre).
The scan
and github
commands now support the --all-organizations
flag. When supplied along with a custom GitHub API URL, Nosey Parker will scan the provided GitHub instance for all organizations to be further enumerated for additional repositories (#126; thank you @seqre).
New rules have been added (thank you @gemesa):
A new generate
command has been added, which generates various assets that are included in prebuilt releases:
Several rules have been fixed that in certain circumstances would fail to match and produce a runtime error message:
The netrc Credentials
rule has been modified to avoid a runtime message about an empty capture group.
The JSON Web Token (base64url-encoded)
rule has been improved to reduce false positives. Thank you @saullocarvalho for the bug report.
The prebuilt releases now include shell completion scripts for bash, fish, elvish, powershell, and zsh, instead of 5 copies of the zsh completions (#132; thank you @Marcool04).
The minimum supported Rust version has been changed from 1.70 to 1.76.
The data model and datastore have been significantly overhauled:
The rules used during scanning are now explicitly recorded in the datastore. Each rule is additionally accompanied by a content-based identifier that uniquely identifies the rule based on its pattern.
Each match is now associated with the rule that produced it, rather than just the rule's name (which can change as rules are modified).
Each match is now assigned a unique content-based identifier.
Findings (i.e., groups of matches with the same capture groups, produced by the same rule) are now represented explicitly in the datastore. Each finding is assigned a unique content-based identifier.
Now, each time a rule matches, a single match object is produced. Each match in the datastore is now associated with an array of capture groups. Previously, a rule whose pattern had multiple capture groups would produce one match object for each group, with each one being associated with a single capture group.
Provenance metadata for blobs is recorded in a much simpler way than before. The new representation explicitly records file and git-based provenance, but also adds explicit support for extensible provenance. This change will make it possible in the future to have Nosey Parker scan and usefully report blobs produced by custom input data enumerators (e.g., a Python script that lists files from the Common Crawl WARC files).
Scores are now associated with matches instead of findings.
Comments can now be associated with both matches and findings, instead of just findings.
The JSON and JSONL report formats have changed. These will stabilize in a future release (#101).
The matching_input
field for matches has been removed and replaced with a new groups
field, which contains an array of base64-encoded bytestrings.
Each match now includes additional rule_text_id
, rule_structural_id
, and structural_id
fields.
The provenance
field of each match is now slightly different.
Schema migration of older Nosey Parker datastores is no longer performed. Previously, this would automatically and silently be done when opening a datastore from an older version. Explicit support for datastore migration may be added back in a future release.
The shell-completions
command has been moved from the top level to a subcommand of generate
.
Published by bradlarsen 11 months ago
A prebuilt multiplatform Docker image for this release is available for x86_64 and ARM64 architectures:
docker pull ghcr.io/praetorian-inc/noseyparker:v0.16.0
Additionally, a prebuilt Alpine-based image is also available for x86_64 and ARM64 architectures:
docker pull ghcr.io/praetorian-inc/noseyparker-alpine:v0.16.0
The scan
command now supports a new --copy-blobs={all,matching,none}
parameter. When specified as matching
, a copy of each encountered blob that has matches will be saved to the datastore's blobs
directory. When specified as all
, a copy of each encountered blob will be saved. The default value is none
. This mechanism exists to aid in ad-hoc downstream investigation. Copied blobs are not used elsewhere in Nosey Parker at this point.
A new advanced global command-line parameter has been exposed:
--sqlite-cache-size=SIZE
to control the pragma cache_size
value used in sqlite database connectionsThe datastore now contains two additional tables for to represent freeform comments and accept/reject status associated with findings. These additional tables are not currently populated in the open-source version of Nosey Parker. The report
command now emits finding status and comment if populated. Note: the datastore format is not settled and is subject to change.
A new "ruleset" mechanism has been added. A ruleset is a named collection of rules that can be selected as a group. The new --ruleset=NAME
parameter to scan
can be used to enable alternative rulesets. Three built-in rulesets are provided (default
, np.assets
and np.hashes
); the special ruleset name all
enables all known rules. See the built-in rulesets at crates/noseyparker/data/default/builtin/rulesets
for an example for writing your own.
The default collection of rules has been pruned down to further emphasize signal-to-noise. Only rules that detect secret things are included in the default collection. Rules that detect other things, such as cloud assets, application IDs, or public keys, are not included in this set. Instead, those are in the np.assets
ruleset, which is not enabled by default. No rules have been removed from Nosey Parker; rather, the defaults have been adjusted to support the most common use case (secrets detection).
Additional checks have been added to the rules check
command:
A new rules list
command is available, which lists available rules and rulesets. This command can emit its output in human-oriented format or in JSON format.
New rules have been added:
A new global --quiet
/ -q
option has been added, which suppresses non-error feedback messages and disables progress bars (#97).
Command-line parameters that can meaningfully accept negative numbers can now be specified without having to use --PARAMETER=NEGATIVE_VALUE
syntax; a space can now separate the paraemter and the value.
Fixed three rules that were missing capture groups:
Due to nuanced details of how scanning is performed, rules without capture groups will never produce reported matches. An additional check was added to the rules check
command and a couple assertions were added that should help prevent this type of error in the future.
Fixed several rules:
The LICENSE, README.md, and CHANGELOG.md files are now included in prebuilt binary releases.
ANSI formatting sequences are now no longer included by default by the report
command when the output is redirected to a file using the -o
/--outfile
parameter (#55).
The scan
command should no longer emit warnings like Failed to decode entry in tree
. These warnings were due to a bug in the Git object parsing code in the gix
dependency, which was fixed upstream.
The rules check
command invocation now behaves differently. It now no longer requires input paths to be specified. It will check the built-in rules for problems, and if additional paths are specified, will check those rules as well. This change was made so that the scan
, rules check
, and rules list
invocations have consistent interfaces.
The default path-based ignore rules in Nosey Parker now ignore packed-refs
files from Git repositories.
Several rules have been changed:
Slack
rule (id np.slack.1
) has been removed, as it was redundant with Slack Token
.Slack Token
has been split into Slack Bot Token
, Slack Legacy Bot Token
, Slack User Token
, and Slack App Token
.CodeClimate
was enhanced to detect additional cases and was renamed to CodeClimate Reporter ID
.md5crypt Hash
(id np.md5.1
) has been renamed to Password Hash (md5crypt)
and re-identified as np.pwhash.1
.bcrypt Hash
(id np.bcrypt.1
) has been renamed to Password Hash (bcrypt)
and re-identified as np.pwhash.2
.Log messages are written to stderr instead of stdout (#97).
Published by bradlarsen about 1 year ago
NOTE: release artifacts take some time to build, and will be added to the release soon
A default value (datastore.np
) is now set for commands that take a datastore parameter (#74). This makes simpler noseyparker
command-line invocations possible.
A new shell-completions
command has been added, which generates shell-specific completion scripts for zsh, bash, fish, powershell, and elvish (#76). These generated completion scripts make discovery of Nosey Parker's command-line API simpler. Thank you @Coruscant11!
The report
command supports a new --max-matches=N
parameter to control the maximum number of matches that will be output for any single finding (#75). A negative number means "no limit".
The scan
command now supports a new --git-history={full,none}
parameter to control whether encountered Git history will be scanned. This defaults to full
, but specifying a value of none
will cause Git history to be ignored.
New rules have been added:
A new disable_tracing
Cargo feature has been added, which disables trace
-level logging and tracing messages. This feature is also aliased by a new release
feature, which is enabled in prebuilt releases.
The NP_LOG
environment variable is inspected at runtime to allow find-grain control over Nosey Parker's diagnostic output. The syntax of this variable are defined by the tracing-subscriber
Rust crate.
All the output formats for the report
command now respect the new --max-matches=N
parameter. Previously, the output formats other than human
would run without limit (i.e., as though --max-matches=-1
had been specified).
The release process is now codified in a shell script: scripts/create-release.zsh
. This emits a release tree at release
in the top-level of the repository, which includes the prebuilt binary as well as shell completions (#80).
The report
command has improved performance when using JSON output format. Previously, the entire JSON output document needed to be accumulated in memory and then written in one step at the end. Now, the JSON output document is written in a streaming fashion, one finding at a time.
mimalloc
is now used as the global allocator (#81). This reduces peak resident memory when scanning large inputs with a high degree of parallelism.
report
command when --format=sarif
is used which caused some metadata to be unintentionally omitted from the output.Published by bradlarsen about 1 year ago
A prebuilt multiplatform Docker image for this release is available for x86_64 and ARM64 architectures:
docker pull ghcr.io/praetorian-inc/noseyparker:v0.14.0
Running noseyparker --version
now emits many compile-time details about the build, which can be useful for troubleshooting (#48).
The github
and scan
commands now support accessing GitHub Enterprise Server instances using the new --github-api-url URL
parameter (#53—thank you @AdnaneKhan!).
New rules have been added:
Rules are now required to have a globally-unique identifier (#62)
Two new advanced global command-line parameters have been exposed:
--rlimit-nofile LIMIT
to control the maximum number of open file descriptors--enable-backtraces BOOL
to control whether backtraces are printed upon panicThe snippet length for matches found by the scan
command can now be controlled with the new --snippet-length BYTES
parameter.
The Git repository cloning behavior in the scan
command can now be controlled with the new --git-clone-mode {mirror,bare}
parameter.
The scan
command now collects additional metadata about blobs. This metadata includes size in bytes and guessed mime type based on filename extension. Optionally, if the non-default libmagic
Cargo feature is enabled, the mime type and charset are guessed by passing the content of the blob through libmagic
(the guts of the file
command-line program).
By default, all this additional metadata is recorded into the datastore for each blob in which matches are found. This can be more precisely controlled using the new --blob-metadata={all,matching,none}
parameter.
This newly-collected metadata is included in output of the report
command.
The scan
command now collects additional metadata about blobs found within Git repositories. Specifically, for each blob found in Git repository history, the set of commits where it was introduced and the accompanying pathname for the blob is collected (#16). This is enabled by default, but can be controlled using the new --git-blob-provenance={first-seen,minimal}
parameter.
This newly-collected metadata is included in output of the report
command.
The datastore schema has been changed in an incompatible way such that migrating existing datastores to the new version is not possible. This was necessary to support the significantly increased metadata that is now collected when scanning. Datastores from earlier releases of Nosey Parker cannot be used with this release; instead, the inputs will have to be rescanned with a new datastore.
The JSON and JSONL output formats for the report
command have changed slightly. In particular, the .matches[].provenance
field is now an array of objects instead of a single object, making it possible to handle situations where a blob is discovered multiple ways. The provenenance
objects have some renamed fields, and contain significantly more metadata than before.
Existing rules were modified to reduce both false positives and false negatives:
The default size of match snippets has been increased from 128 bytes before and after to 256.
This typically gives 4-7 lines of context before and after each match.
When a Git repository is cloned, the default behavior is to match git clone --bare
instead of git clone --mirror
. This new default behavior results in cloning potentially less content, but avoids cloning content from forks from repositories hosted on GitHub.
The command-line help has been refined for clarity.
Scanning performance has been improved on particular workloads by as much as 2x by recording matches to the datastore in larger batches. This is particularly relevant to heavy multithreaded scanning workloads where the inputs have many matches.
Python is no longer required as a build-time dependency for vectorscan-sys
.
A typo was fixed in the Okta API Key rule that caused it to truncate the secret.
The scan
command now correctly reports the number of newly-seen matches when reusing an existing datastore.
Published by bradlarsen over 1 year ago
A prebuilt multiplatform Docker image for this release is available for x86_64 and ARM64 architectures:
docker pull ghcr.io/praetorian-inc/noseyparker:v0.13.0
Nosey Parker now statically links against a bundled version of Vectorscan for regular expression matching instead of Hyperscan (#5). This makes building from source simpler, particularly for ARM-based platforms. This also simplifies distribution, as a precompiled noseyparker
binary now has no runtime library dependencies on non-default libraries.
Several existing rules were modified to reduce false positives and false negatives:
New rules have been added:
References have been added for several rules:
git
binary installed. Previously this was missing, causing the scan
command to fail when the --git-url
, --github-user
, or --github-organization
input specifiers were used (#38).Published by bradlarsen over 1 year ago
A prebuilt Docker image for this release is available for x86_64 architectures:
docker pull ghcr.io/praetorian-inc/noseyparker:v0.12.0
The scan
command can now be given Git https URLs, GitHub usernames, and GitHub organization names as inputs, and will enumerate, clone, and scan as appropriate (#14).
Nosey Parker now has rudimentary support for enumerating repositories from GitHub users and organizations (#15). The new github repos list
command uses the GitHub REST API to enumerate repositories belonging to one or more users or organizations. An optional GitHub Personal Access Token can be provided via the NP_GITHUB_TOKEN
environment variable.
Nosey Parker now has an optional rule_profiling
crate feature that causes performance-related statistics to be collected and reported when scanning. This feature imposes some performance cost and is only useful to rule authors, and so is disabled by default.
Many new rules have been added:
These rules match token formats that are well-specified fixed-length strings with notable prefixes or suffixes, and so should produce very few false positives.
Several existing rules were modified to improve signal-to-noise:
The report
command now offers rudimentary SARIF support (#4). Thanks you @Coruscant11!
Several default rules have been revised to improve performance of the matching engine and to produce fewer false positives. In particular, several rules previously had avoided using a trailing \b
anchor after secret content which could include a literal -
character, due to a matching discrepancy between Hyperscan and Rust's regex
library. These have been revised to use a more complicated but functional anchoring pattern.
The JSON Web Token (base64url-encoded)
rule has been changed to only produce a single match group instead of three.
The Google Client Secret
rule has been improved to detect additional occurrences and has been renamed to Google OAuth Client Secret
.
Blobs are now deduplicated at enumeration time when first enumerating a Git repository, rather than only at scan time. This results in more accurate progress bars.
When scanning, Git repositories are now opened twice: once at input enumeration time, and once at scanning time. This drastically reduces the amount of memory required to scan a large number of Git repositories.
Published by bradlarsen almost 2 years ago
This is the first released version of Nosey Parker. Its scan
, summarize
, and report
commands are functional. It is able to scan files, directories, and the complete history of Git repositories at several hundred megabytes per second per core. It comes with 58 rules.
A prebuilt Docker image for this release is available for x86_64 architectures:
docker pull ghcr.io/praetorian-inc/noseyparker:v0.11.0