Analyst Data
Analyst Data is a collaborative annotation layer that lets analysts enrich anything in MISP — an event, an attribute, an object, an event report, a galaxy cluster and more — without modifying the annotated element itself. Instead of editing another organisation’s event (which you cannot do) or spawning a proposal, you attach your own record alongside it. Each record has its own identity, its own distribution, and its own author, so it can be shared and synchronised independently of the thing it describes.
The examples describe MISP 3.0.0 from the 3.0_wip development branch. The 3.0 interface surfaces event analysis in Assessment and value-level analysis in Value Intelligence → Collaboration. Select a screenshot to enlarge it.
There are three types of analyst data:
- Note — free-text commentary, with a language tag.
- Opinion — a numeric agreement score from 0 to 100 plus an optional justification.
- Relationship — a directed, labelled link from one element to another.
Because each record is identified by its own UUID and carries its creator organisation, analyst data is inherently federated: notes, opinions and relationships propagate between instances through synchronisation just like events do, letting a whole community build a shared, attributable body of analysis around the same intelligence.
Creating or editing analyst data requires a role with the Analyst Data Creator permission (perm_analyst_data). Viewing analyst data only requires being logged in (subject to the record’s distribution).
The three types
Notes
A Note is a piece of free-text commentary attached to a target element. Beyond the note body itself, a note carries a language (an RFC 5646 tag such as en, fr, de) so that multilingual communities can annotate in their own language and filter accordingly.
Use notes to record analysis, caveats, remediation guidance, or any context that does not fit neatly into an attribute or tag.
Opinions
An Opinion expresses how much you agree with the element it is attached to, as a score from 0 to 100, with an optional free-text comment explaining your reasoning. The scale is read as five bands:
| Score | Meaning |
|---|---|
| 0–20 | Strongly disagree |
| 21–40 | Disagree |
| 41–60 | Neutral (50 = neutral midpoint) |
| 61–80 | Agree |
| 81–100 | Strongly agree |
In the interface the score is shown as a coloured gradient bar with the numeric value. Opinions are the natural way to signal (dis)agreement with a shared indicator or assessment without altering it — for example, to flag a suspected false positive or to endorse a high-value indicator.
Relationships
A Relationship is a directed edge from a source element to a related element, carrying a free-text relationship type label (for example similar-to, derived-from, part-of). The relationship type is a free-text field with autocomplete — the suggestions combine relationship types already used in analyst data with MISP’s object-reference vocabulary — so you are not restricted to a fixed list.
A relationship can link any two supported elements, even across events and across instances, and targets are normally referenced by UUID. In MISP 3.0, either end can instead target an attribute value using its SHA-256 address (see Value-level analyst data). A record cannot relate an element to itself.
Analyst-data Relationships are not the same as two other MISP concepts that also link things together:
- MISP object references link an attribute or object to another object within the same event; they are part of the event’s own data model, not analyst data.
- Galaxy cluster relationships link galaxy clusters to one another as part of the galaxy data model.
Only analyst-data Relationships are individually authored by an organisation, individually distributable, blocklistable, and synchronised through the analyst-data pipeline. See MISP Objects and Galaxies for the other two.
What can be annotated
Analyst data can be attached to any of the following target types: Attribute, Event, Event Report, Galaxy Cluster, Galaxy, Object, Organisation, Sharing Group, and — because analyst data can annotate analyst data — Note, Opinion and Relationship themselves. Attaching a note or opinion to another note produces a thread, so a discussion can build up around a single element. The MISP 3.0 development branch also accepts Collection, Workspace, and Attribute Value (AttributeValue) targets. An attribute value identifies the value across events rather than one attribute occurrence.
In practice, notes, opinions and relationships are surfaced automatically when you view an event, attribute, object, event report or galaxy cluster; the other target types are valid anchors but are not all rendered inline in their own views. MISP 3.0 surfaces value-level analysis in Value Intelligence; see the workflow below.
Working with analyst data in the UI
Open an event’s General tab and inspect its Assessment card. Its add control offers the annotation types permitted for your account; attribute and object rows have their own analysis controls. Use Data points → Analyst Data for the standalone list. The form identifies the target and asks for distribution plus the type-specific content.
Open an annotation to inspect its own identity, author, target, audience, and available editing or thread controls. An Opinion score is an assessment in context, not a probability that an indicator is malicious.
Opening a record shows it together with its thread — nested notes and opinions — and its relationships, split into:
- Outbound relationships (edges this element points out to), and
- Inbound relationships (edges from other elements that point at this element).
Inbound relationships are computed automatically, so you can see who has linked to an element even if the link was created elsewhere.
Value-level analyst data in MISP 3.0
In the 3.0_wip development interface, open Data points → Value Intelligence, look up the value, and select Open intelligence. The Collaboration tab offers Add note and Add opinion for the value itself. Use the attribute occurrence instead when your statement depends on a particular source, event, or observation window. See Best Practices for a worked example and screenshots.
A value has no UUID. Its address is the lowercase hexadecimal SHA-256 of its stored text, encoded as UTF-8. Compute it from the value returned by MISP after normalization, rather than from unprocessed input. Notes, Opinions, and either end of a Relationship can use this AttributeValue target. The analyst-data record itself still has its own UUID.
Set the annotation’s distribution separately from its target. An annotation’s own sharing rules determine who may read it; access to the value’s holder event does not restrict text you put in a more broadly shared Note. The value reference resolves to text only when the reader can see at least one attribute holding it. The value text does not travel with the reference. Value-addressed analyst data is withheld from MISP peers older than 3.0 during synchronization.
Distribution and sharing
Every analyst-data record has its own distribution setting, independent of the element it annotates, using the same levels as events:
| Level | Meaning |
|---|---|
| 0 | Your organisation only |
| 1 | This community only |
| 2 | Connected communities |
| 3 | All communities |
| 4 | Sharing group (an explicit list of organisations) |
The default level for new analyst data is controlled by the Plugin.default_analyst_data_distribution server setting (default: This community only). When the level is Sharing group, MISP validates that you are allowed to use the chosen sharing group. Visibility follows the usual rules: you can see a record if your organisation created it, if it is distributed to your community or wider, or if it is in a sharing group you belong to.
Collaboration and synchronisation
Analyst data is built for federated collaboration:
- Identity — each record has a globally unique UUID that is stable across instances, so the same note is recognised everywhere it travels.
- Authorship — every record records its creator organisation; only the creating organisation (or a site admin) can edit or delete it. Other organisations add their own parallel notes and opinions rather than changing yours.
- Synchronisation — analyst data is pushed and pulled between instances alongside events, provided both sides support the feature and the per-server Push analyst data / Pull analyst data toggles are enabled (see Sharing). Distribution and server filters determine what is exchanged. Community-limited distribution can be reduced at a synchronization boundary; do not assume that All communities or a sharing group is decremented like a numeric hop count. Records received from a remote instance are marked as locked to protect the source copy from unintended local edits.
Analyst Data Blocklist
Because analyst data is synchronised, simply deleting an unwanted record locally would not stop it being pulled back in from a remote instance. To prevent that, MISP keeps an Analyst Data Blocklist keyed on the record UUID. When a record is (hard-)deleted, a blocklist entry is created automatically — recording the UUID, a short human-readable summary, and the creator organisation — so the record is ignored if it is offered again during synchronisation.
The blocklist is managed by site administrators under the administration menus and behaves like the other blocklists in MISP.
API and permissions
Analyst data is fully accessible over the REST API under the /analystData/ controller. The main actions are:
| Action | Purpose | Permission |
|---|---|---|
add/{type}/{object_address}/{object_type} |
Create a note/opinion/relationship on a target | perm_add + perm_analyst_data |
edit/{type}/{id or uuid} |
Modify a record you own | perm_add + perm_analyst_data |
delete/{type}/{id or uuid} |
Delete a record you own | perm_add + perm_analyst_data |
view/{type}/{id or uuid} |
View a single record | any logged-in user |
index/{type} |
List records of a type | any logged-in user |
getChildren/{type}/{uuid} |
Fetch a record’s thread | any logged-in user |
where {type} is Note, Opinion or Relationship (or all, letting MISP deduce the type from the UUID). The synchronisation endpoints used between instances require the perm_sync permission. As with the rest of the API, requests and responses use the standard MISP JSON envelope — see the Automation & API chapter.
Value targets through the 3.0 API
For a value target, use its SHA-256 address as {object_address} and AttributeValue as {object_type}. These endpoints use the same analyst-data permissions and distribution checks as UUID targets:
| Endpoint | Purpose |
|---|---|
POST /analystData/add/Note/{sha256}/AttributeValue |
Add a Note to the value. |
POST /analystData/add/Opinion/{sha256}/AttributeValue |
Add an Opinion to the value. |
POST /analystData/add/Relationship/{sha256}/AttributeValue |
Create a Relationship whose source is the value. |
GET /analystData/viewForObject/AttributeValue/{sha256} |
Retrieve visible Notes, Opinions, and Relationships for the value. |
For example, the Note endpoint accepts:
{
"Note": {
"note": "Reserved documentation value; do not use for operational detection.",
"language": "en",
"distribution": 0
}
}For a Relationship pointing to a value, set related_object_type to AttributeValue and supply related_object_sha256 instead of related_object_uuid. The source and destination can independently use a value address or a UUID target.

