Skip to main content

Ochrona checks your open source dependencies for vulnerabilities and policy violations.

Project description

Ochrona PyPI Versions License Downloads Vuln DB Version codecov Code style: black


Ochrona is a free solution for securing the dependencies used in Python projects. Ochrona also includes support for policies which give you additional control over what aspects of your dependency usage you'd like to be alerted on.

The Ochrona maintainers care deeply about Developer Experience (DX), if you have any feedback or run into issues please open an issue here.

Vulnerability Data

Ochrona maintains its own database of vulnerabilities impacting Python packages. You're welcome to check out the database here. This database is updated frequently using data from NVD, Github, and other sources. Ochrona will update its local copy of the database if a new version exists.

Supported file types

  • *requirements*.txt
  • Pipfile.lock
  • poetry.lock
  • *constraints*.txt
  • environment.yml
  • tox.ini


via pip

pip install ochrona

via pipenv

pipenv install --dev ochrona

via poetry

poetry add -D ochrona


via command line args

Arg Description Type Example
--dir Directory to recursively search for dependencies files to scan [.] path /User/me/my_project
--exclude_dir Directory names that should be excluded from recursive search. Comma separated str build,dev
--file Single dependency file to scan file /User/me/my_project/requirements.txt
--debug Enable debug logging [False] bool True
--silent Silent mode [False] bool True
--report_type The report type that's desired [BASIC] str XML
--output Location for report output path /User/me/my_project/logs
--exit Exit with Code 0 regardless of vulnerability findings. [False] bool True
--ignore Ignore a CVE or package str requests
--include_dev Include develop dependencies from Pipfile.lock [False] bool True

via environment variables

Variable Name Corresponding Arg

via .ochrona.yml

There is an empty .ochrona.yml file included in the repo. | Key | Description | Type | Example | |-|-|-|-| | dir | Directory to recursively search for dependencies files to scan [.] | path | /User/me/my_project | | exclude_dir | Directory names that should be excluded from recursive search. | list | build | | file | Single dependency file to scan | file | /User/me/my_project/requirements.txt | | debug | Enable debug logging [false] | bool | true | | silent | Silent mode [false] | bool | true | | report_type | The report type that's desired [BASIC] | str | XML | | report_location | Location for report output [.] | path | /User/me/my_project/logs | | exit | Exit with Code 0 regardless of vulnerability findings [false] | bool | true | | ignore | Ignore a CVE or package name | str | requests | | include_dev | Include develop dependencies from files that support dev/required dependencies [false] | bool | true | | color_output | Whether or not std out text should use color. Note: this is enabled by default when running in a non-Windows environment [true] | bool | false | | policies | Policies are a way of defining additional checks against your dependencies. See here for more details | array | details |


# debug: true
# silent: false
# dir: .
# report_type: JSON
# report_location: .
# ignore: requests
# include_dev: false
# color_output: false
# policies:
#  - license_type NIN APSL,GPL-PA,JSON


Policies are a way to add additional check to your Python dependency usage. Policies can be defined using conditional and logical syntax. These generic policy definitions allow you to define unique, custom policies to fit your need, and they can be extensible as new fields and capabilities are added.

Policy vioations are not the same as vulnerabilities, however violations will cause Ochrona to emit a failure exit code and the output will include details about the policy violation.

Policy Syntax

At their most basic, policies are defined using conditional statements and logical operators. A conditional statement is structured as <field><operator><value>, for example, license_type == MIT. Whitespace is always ignored during policy evaluation.

Allowed Fields

Name Description
name The name of a package.
license_type An (SPDX)[] license type for a package.
latest_version The most recent version of a package.
latest_update The timestamp for when a package was last updated. ISO-8601 Format YYYY-MM-DDTHH:MM-SS.ffffffZ
release_count The number of releases a package has on Pypi.

Allowed Conditional Operators

Operator Description
== An equals operator for comparing exact string matches.
!= A NOT equals operators for non-matching strings.
< Less than, for comparing numerical or string values.
<= Less than or equal to.
> Greater than, for comparing numerical or string values.
>= Greater than or equal to.
IN For checking whether a value exists within a set.
NIN For checking that a value does not exist within a set.

Allowed Logical Operators

Operator Description
AND For checking that all conditions are true.
OR For checking that at least one condition is true.

Special Values

Value Description
NOW-N Shorthand for an ISO 8601 formatted date in the past. N will be an integer number of days.

Policy Examples

# Policy to check that a license type is in my aproved list
license_type IN MIT,ISC,Apache-2.0

# Policy to check that all packages have been updated this year
latest_update >= NOW-365

Legacy Policies

Legacy policies can also be defined using their name and one or more conditions which are evaluated at scan time. Legacy policies will be removed in a future release and you are encouraged to use generic policies for all new policies.

For example, the license_type policy allows you to be alerted if one of your dependency's open-source license is not part of your "Allow List" or if it uses a license from your "Deny List".

Legacy Policy Types

Name Description Fields
package_name Allows for checking whether the dependencies used are all from an allow_list or contain any values from a deny_list. You may define allow_list or deny_list, but not both. Field values should be defined as a comma-separated string. allow_list, deny_list
license_type Allows for checking whether the licenses of dependencies used are all from an allow_list or contain any values from a deny_list. You may define allow_list or deny_list, but not both. Field values should be defined as a comma-separated string. (SPDX)[] license ids should be used. allow_list, deny_list

Usage Examples

Default Mode

$ ochrona 

This will search for any supported dependency files recursively from the run location. It will output rules in the BASIC format to stdout. The program will exit with an error exit code if any confirmed vulnerabilities are found.

Standard error code with Junit XML reporting saved to file

$ ochrona --exit --report_type XML --output ./output

Safe Import Mode

In this mode ochrona acts as a safe wrapper around standard pip installs to ensure that a package and it's dependencies are safe before installing. This action preemptively checks a package and only imports if no vulnerabilities are found. It can be used with a base package (i.e. requests), or with a package pinned to an exact version (i.e. requests==2.21.0). It also supports importing a requirements.txt style, the pip equivalent of pip install -r <file>.

$ ochrona --install <package_name>|<requirements.txt>

stdin Support

Ochrona supports supplying dependencies via stdin and can accept a PEP-508 complaint (i.e. requirements.txt) formated string, or a single dependency. Single dependencies can be supplied as the first argument or piped.

Single dependency via stdin

$ ochrona urllib3==1.26.4
$ echo "urllib3==1.26.4" | ochrona

Multi-dependency via stdin

$ pip freeze | ochrona
$ pipenv lock -r | ochrona
$ cat requirements.txt | ochrona

Docker Support

Ochrona can be run via Docker. This is useful for the paranoid who may worry that an installed module could have modified the Python package namespace and allow malicious packages to bypass Ochrona's security checks. We've added this support in response to CVE-2020-5252 which was disclosed prior to Ochrona and affects several other similar tools.

Dockerized Ochrona

$ pip freeze | docker run -i --rm ochrona/ochrona ochrona

Dockerized Ochrona with environment variables

$ pip freeze | docker run -i -e OCHRONA_IGNORED_VULNS=requests --rm ochrona/ochrona ochrona

Output Formats

Ochrona supports several built in output options include a BASIC and FULL plaintext reports, as well as a Junit style XML report or a JSON style report for incorporating with other tools.



XML (Junit)




Let the world know you're keeping your project safe with Ochrona. Add our shield to your by adding the following line.


Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for ochrona, version 1.1.0
Filename, size File type Python version Upload date Hashes
Filename, size ochrona-1.1.0-py3-none-any.whl (43.1 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size ochrona-1.1.0.tar.gz (33.8 kB) File type Source Python version None Upload date Hashes View

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page