Pull the indicators out of the noise, and package them PARCEL
The end of an investigation is a pile of indicators buried in notes and logs. This finds them, sorts them, drops the noise, defangs the dangerous ones, and writes them in a format another team can ingest.
The text
Paste a report, a log, notes, or SPECIMEN's output. Defanged indicators (hxxp, [.]) are understood.
Indicators
Untick anything benign before you export. Shown defanged; the STIX export keeps the real values.
How it works
What it does, and does not
It finds things shaped like indicators, dedupes them and drops private ranges and documentation domains; it does not judge whether each is malicious, so review before you share. Defanging keeps a malicious URL from going live in a ticket; the STIX bundle carries the real values for a platform to ingest. Everything stays in the browser. Pairs with SPECIMEN and MITRE.
Questions people ask about PARCEL
What is an indicator of compromise?
An observable artefact that points to an intrusion and can be searched for elsewhere: a file hash, a domain, an IP address, a URL, an email address, a file path or a registry key. Its value is that another team can look for the same thing in their own logs, mail and endpoints, or block it.
A CVE ID is not one. It names a vulnerability, not evidence that anybody used it. PARCEL lists CVEs because a report often needs them as context, but it leaves them out of the STIX bundle.
What does defanging an indicator mean?
Rewriting it so that it cannot be clicked or resolved by accident when it is pasted into a ticket, a chat or an email, while a person can still read it. https://evil.com/x becomes hxxps[://]evil[.]com/x.
PARCEL defangs URLs and domains this way, brackets the dots and colons of IP addresses, and brackets the @ and dots of email addresses. Hashes, paths, registry keys and CVEs are left as they are, because they cannot go live. Defanged input is read too: hxxp, [.], (.), [dot], [@], [at] and [:] are all undone before searching.
What is in the STIX 2.1 bundle?
One STIX indicator object for each ticked hash, URL, domain, IPv4 address, IPv6 address and email address, with the real, refanged value in a STIX pattern such as [domain-name:value = 'evil.com'], the label malicious-activity, and the moment of export as its creation and valid_from time.
CVEs, registry keys and paths are not in the bundle. They are in the copied list and the CSV, whose columns are type, indicator and defanged. There is no MISP export.
Why are some IP addresses and domains left out?
To keep a report free of noise. IPv4 addresses in 10.0.0.0/8, 127.0.0.0/8, 0.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16, 169.254.0.0/16 and everything from 224.0.0.0 up are dropped, with the shared carrier NAT range 100.64.0.0/10 and the documentation ranges 192.0.2.0/24, 198.51.100.0/24 and 203.0.113.0/24. So are the IPv6 loopback ::1, link-local fe80::/10, unique local fc00::/7 and documentation 2001:db8::/32 addresses.
Domains under example.com, example.net, example.org, localhost, .test, .invalid, .local and .arpa are dropped, and so is any name ending in a common file extension, such as .exe, .dll, .php or .zip, because it is far more often a file name. A real domain on the .zip or .sh top-level domain is dropped with them, so check the source text if one matters.
How does it tell an IPv6 address from a time or a MAC address?
By its shape. An IPv6 address is eight groups of one to four hexadecimal digits separated by colons, or fewer with a single :: standing for the groups of zeros left out, so 2600:1f18:4a3::7 is one. A time of day such as 10:22:33 has three groups and no ::, and a MAC address such as aa:bb:cc:dd:ee:ff has six, so neither is listed.
Addresses are lower-cased so the same one written in two cases is listed once. A colon that only ends a sentence is left off.
How does it tell an MD5 from a SHA-1 or a SHA-256?
By length. A run of exactly 32 hexadecimal characters is listed as MD5, 40 as SHA-1 and 64 as SHA-256, and the run has to stand on its own, so part of a SHA-256 is never reported as an MD5. Hashes are lower-cased so the same hash in two cases is listed once.
Length is all it knows. Any other identifier that happens to be 32 hexadecimal characters will be listed as an MD5.
Can I take things out before I share the list?
Yes. Every indicator starts ticked; untick the ones that are benign, such as your own domains or a vendor’s update server, and they are left out of the copied list, the CSV and the STIX bundle alike. PARCEL finds things shaped like indicators and does not judge whether any of them is malicious, so this review is the step that makes the list worth sharing.
Doing this often? The Sovereign Console pulls every indicator across a whole case into one place, and exports a self-contained report. On its own, US$2 a month; with everything else, US$5 a month or US$48 a year.