Automation

Keep unwanted cases out, add repeats to the open case they repeat, and act on cases when they open or change.

Overview

Each new case goes through three steps, in this order:

  1. Drop rules refuse the case before DFIRe creates it. The first rule that matches drops the case, and no later step runs. See Drop rules.
  2. Deduplication, when it is on, compares the case with the open cases. If it repeats one, DFIRe stores it in that case's duplicate list without creating another case. See Deduplication.
  3. Automation rules run on the case that DFIRe created, and again when the case changes. See How an automation rule is built.

All three live under Settings → Automation, which shows them in the order they run. Only a superuser opens it. See Access control.

A typical automation rule keeps the right responders on a case without anyone remembering to add them. For example, it invites the on-call network engineer to the Slack channel of every case in the Product Security project, or assigns a team lead based on the playbook. Slack invites work for people who have no DFIRe account, so external responders join the conversation as soon as the case exists.

The three steps also handle an integration that opens one case per alert. A new Linux server, for example, can raise a hundred CIS benchmark alerts at once. Each has its own title, such as Center for Internet Security Debian Linux 12 Benchmark v1.1.0: Ensure the audit configuration is immutable. on http-proxy. One automation rule closes every such alert, from any machine and at any level. It needs Title contains set to Center for Internet Security Debian Linux 12 Benchmark and the action Close case.

A drop rule with the same condition keeps the alerts out and leaves no record. Deduplication on Case title match stores an alert that repeats an open case in its duplicate list.

Drop rules

A drop rule refuses a new case that matches its conditions. A dropped case leaves nothing behind: no case, no case number, no timeline entry, no audit log entry, no Slack channel, no webhook and no automation run.

DFIRe checks the enabled drop rules from top to bottom, and the first rule that matches drops the case. The rules apply to every new case: from the New Case dialog, the API, the MCP server, the Slack form and case import. They apply to superusers too. See What the creator sees.

  1. Open Settings → Automation and click Add drop rule

    Give the rule a name and, if you like, a description. See Rule names and descriptions. Use the Enabled switch to turn a rule off without deleting it.

  2. Set the conditions

    The rule drops a case that meets every condition you set. See Conditions. A drop rule needs at least one condition, because a rule without one would drop every new case. DFIRe refuses to save it.

  3. Save the rule

    A new rule joins the end of the list. Drag a rule to change its place in the order, and use the switch on its row to turn it on or off.

Each row shows how many cases the rule has dropped and when it last matched. DFIRe keeps these counts in its cache, not in the database, and clearing the cache resets them. An import preview does not count as a match.

Deduplication

Deduplication stores a repeated alert in the open case's Duplicates list. It takes no case number and adds no timeline entry. Deduplication runs after the drop rules, so it never sees a dropped case.

It is one setting for the whole installation. Under Settings → Automation, turn on Deduplicate new cases, choose the criteria and select Save. A new case repeats an open case when every selected criterion matches:

Criterion The open case matches when
Case title matchIts title is the title of the new case, letter case included.
Case tags matchIt carries exactly the same tags as the new case, in any order: no fewer and no more. A new case without tags never matches on this criterion.
Matching attribute values for keysFor each attribute key you add, it holds the same value as the new case. A new case without one of the keys never matches. Use the attribute's stored key, not the label a form shows. The key field suggests the keys your playbooks define, and you can type any other key.
Playbook match, Project match, Mode match, Severity / priority matchThat field is the same as on the new case.

Created within the last limits the comparison to open cases created in that many hours, from 1 to 8760. Leave it empty for no time limit. To turn deduplication on, choose at least one criterion besides the time window. The window alone would add every new case to the most recent open case.

Only open cases take part. Closed and archived cases never match. When several open cases match, the most recently created one receives the duplicate.

Choose the criteria with care. Tags alone are rarely enough because unrelated cases can carry the same tags. Combine tags with the title, or with an attribute that identifies the alert, such as a rule ID.

Stored duplicates

A duplicate keeps the alert's description, fields, tags, evidence items and manual timeline entries. Its items and entries remain stored data until someone makes it a case. The record also shows when it arrived, who sent it and the interface used. It sends no Slack post, timeline webhook or notification. The audit log records the receipt on the original case.

Each receipt raises the original case's duplicate count and moves its Updated time. The Dashboard shows the count in its Duplicates column and card badge. The case header and Case Overview show it too. Deleting one duplicate, promoting or pruning stored records keeps this count. Clearing unlinked records resets it to the linked records kept, or zero.

Keep the last N duplicates per case defaults to 100 and accepts any whole number of 1 or more. This limit counts ordinary, unpromoted duplicates. Each new duplicate removes the oldest ordinary rows beyond the limit. DFIRe keeps records linked to created cases and snapshots made by manual deduplication outside this limit.

The sender needs no access to the original case. If someone closes or deletes it before DFIRe records the repeat, DFIRe creates a new case.

Open the case's Duplicates tab and select a record to read its details. Create case restores its fields, items and manual entries into a new case. Promotion skips drop rules and deduplication, and the duplicate stays linked to the case it created. Delete duplicates without a case removes unlinked records and keeps these links. See Duplicates.

Case import always creates a separate case and restores any stored duplicates in its export. Drop rules still apply to the imported case.

Escalating a case that keeps receiving repeats

An automation rule with the X duplicates received trigger runs when the count reaches the number of duplicates you set. Combine it with a Set severity / priority action to raise an alert that repeats. For example, the rule 20 duplicates received with Set severity / priority to High raises a low severity case to high on its twentieth repeat. After clearing stored duplicates, new repeats can trigger the rule again when the count reaches its threshold.

If Severity / priority match is one of the deduplication criteria, a case whose severity the rule raised no longer matches the repeats, which arrive with the old severity. The next repeat then opens a new case. To keep every repeat on one case, leave Severity / priority match out of the criteria.

While deduplication is on, DFIRe creates new cases one at a time, so it never creates two identical cases that arrive together. The setting shows how many matches it has had and when the last one was. Like the drop rule counts, clearing the cache resets these figures.

What the creator sees

Whoever creates the case learns what happened to it, and the answer names the drop rule or the open case. An automation rule closes a case only after DFIRe has created it, so the creator of such a case gets the normal answer.

Where the case comes from Dropped Repeat of an open case
New Case dialog The message Not created. The case matches the drop rule "CIS benchmark noise". The message Not created. Added to CASE-2026-640 as a duplicate. It has an Open case button when you may open that case.
Slack /dfire create The same messages, shown only to you.
API POST /api/cases/ 200 with dropped set to true and drop_rule, the rule's id and name. 200 with deduplicated: true, duplicate_id, case_id and case_number. case holds the original when the caller may read it.
MCP create_case The same fields as the API.
Case import The import dialog and its preview name the rule, and DFIRe imports nothing. The API answers with dropped and drop_rule as above, and the import summary names the rule under dropped. Never deduplicated.

A created case answers 201 with the case. The other two answers carry no id, so an API client that reads the new case's id checks dropped and deduplicated first.

How an automation rule is built

An automation rule has three parts: one or more triggers (things that happen to a case), optional conditions the case must meet, and one or more actions (what DFIRe does). When a case opens or changes, each enabled rule whose trigger and conditions match runs its actions.

The rule editor reads top to bottom as a sentence, Trigger when … Only if the case matches these conditions … Then do this …. It shows a one-line preview of the rule as you build it.

  1. Open Settings → Automation and click Add automation rule

    Give the rule a name and, if you like, a description. See Rule names and descriptions. Use the Enabled switch to turn a rule off without deleting it.

  2. Choose one or more triggers

    The rule runs when any of its triggers occurs. See Triggers.

  3. Narrow with conditions (optional)

    Limit the rule to cases with a mode, severity or priority, playbook, project, tags, title, attributes or creator. Leave a condition empty, or on Any, to match all cases. See Conditions.

  4. Add one or more actions

    Choose what happens when the rule matches. See Actions.

  5. Save the rule

Execution order

Automation rules run in the order they appear on the Automation settings page, top to bottom. Drag a rule by its handle to change its position. The number next to each rule is its place in the sequence. A new rule joins the end of the list.

When several rules match the same case change, they run one after another in this order, so a later rule can override what an earlier one did. This makes default-and-exception setups predictable: put a rule that assigns your default case lead first, and a rule that assigns a different lead to critical cases after it. For a critical case both run, and the later rule's lead wins. A single case edit runs all its matching rules as one sequence. That holds even when the edit changes several things at once, such as severity and project in the same save.

Close case, Archive case and Stop processing further rules end the sequence. Once one of them has run, no later rule runs on the case for that change. See Closing and archiving a case.

Triggers

A trigger is the case event that starts the rule. Select one or more:

Trigger Fires when
Case opened A case is created. Covers both investigations and incidents. Narrow to one with the Case mode condition.
Escalated to incident An investigation is escalated to incident mode.
Severity / priority escalated A case's severity or priority is raised (for example, medium to high). A change to or from undetermined does not count.
Severity / priority downgraded A case's severity or priority is lowered. A change to or from undetermined does not count.
Playbook changed A case's playbook is changed. Conditions match the new playbook.
Project changed A case is moved to a different project. Conditions match the new project.
X duplicates received A case receives its duplicate number X from deduplication. Enter the number in Duplicates, 1 or more. The rule runs when the count reaches that number. Clearing unlinked duplicates resets the count, so new repeats can trigger the rule again. See Escalating a case that keeps receiving repeats.

Conditions

Conditions select the cases a rule applies to. Drop rules and automation rules use the same conditions. A rule matches only when the case meets every condition you set. Within one condition, matching any of the chosen values is enough, except for tags and attributes, where every listed one must hold. A condition left empty, or on Any, places no restriction.

Condition Holds when
Case mode isThe case is an investigation or an incident, as chosen.
Severity / priority isThe case is critical, high, medium, low, info or undetermined, as chosen.
Playbook isThe case has one of the chosen playbooks (case types).
Project isThe case is in one of the chosen projects.
Has these tagsEvery listed tag is on the case. The comparison is exact and case-sensitive. The Slack form sets no tags, so this condition does not match a case as the form creates it.
Title containsThe title contains the text, ignoring letter case. With Regular expression on, the text is a pattern that matches anywhere in the title. See Title text and regular expressions.
Has these attributesEvery listed key = value pair holds. See Attribute conditions.
Creator isOne of the chosen accounts created the case. The list offers the active accounts, including the service accounts that integrations use. An inactive account stays in the list while the rule names it, so you can remove it.

A drop rule reads the new case before DFIRe stores it. An automation rule reads the case after the change. The automation rule editor greys out combinations that can never occur for the chosen triggers. For example, with Severity / priority escalated selected, the info level is unavailable, because escalation can never leave a case at the lowest level. Undetermined is unavailable with both severity / priority triggers and with Escalated to incident. A change to or from it is neither a raise nor a drop, and an incident always has a severity.

Title text and regular expressions

By default, Title contains looks for the text as typed, so permissions (owner) finds those words with their brackets. Turn on Regular expression to write a pattern instead. It also ignores letter case and matches anywhere in the title. Start it with ^ to match only titles that begin with the text, and end it with $ to match only titles that end with it.

DFIRe checks the pattern when you save the rule, and refuses one that can take unlimited time to match. Drop rules run while DFIRe creates the case, so such a pattern would hold up every new case. The check follows the rules of a webhook's content filter. A pattern can be up to 500 characters long, and DFIRe refuses these shapes:

  • An inline flag, such as (?i) or (?x).
  • A repeated group whose contents repeat or offer alternatives, such as (a+)+ or (a|aa)+.
  • Two repeats written next to each other, such as a*a*b.
  • More than two open-ended repeats, such as * or +, or more than one in a pattern that does not start with ^. DFIRe refuses CIS.*Debian.*12 and accepts ^CIS.*Debian.*12. The ^ must stand outside any brackets, so put | alternatives in a group after it, as in ^(CIS|NIST).*Debian.*12.

The check reads the shape of the pattern, so it also refuses some patterns that would finish quickly. For most titles, the plain text is enough.

Attribute conditions

Each attribute condition is a key and a value. Use the attribute's stored key, not the label a form shows. The key field suggests the keys your playbooks define. You can also type a key that no playbook defines, because an integration can set attributes that the playbook does not define. A case without the key does not match.

A stored text value must be exactly the value you type, letter case included. Any other value compares as its JSON text, so type true, 5 or null.

When a chosen playbook, project or account is deleted

A condition keeps the playbooks, projects and accounts you chose. When someone deletes one of them, the condition can no longer match any case. DFIRe then turns off every enabled drop rule and automation rule that names it, and the rule shows why. The audit log records the change. The editor shows the deleted value as Unknown playbook, Unknown project or Unknown user. Remove it or choose another value, then turn the rule back on. Saving the rule or turning it on removes the message.

Actions

Actions are what the rule does when it matches. A rule can carry several, and runs them in the order they are listed. Inviting someone already in a channel, or adding someone already on the team, does nothing, so the team and Slack actions are safe to repeat. Close case, Archive case and Stop processing further rules end the run instead. Each must be the last action of its rule, and a rule can hold only one of them.

Action Effect
Invite Slack user to case channel Adds a Slack user to the case's Slack channel. Works for people without a DFIRe account.
Remove Slack user from case channel Removes a Slack user from the case's Slack channel.
Add user to case team Adds a DFIRe user to the case team as lead, investigator, or viewer.
Remove user from case team Removes a named DFIRe user from the case team (and from the Slack channel).
Add group to case team Adds the members of a role group to the case team as investigators or viewers. See Adding and removing a group.
Remove group from case team Removes every member of a role group from the case team (and from the Slack channel).
Fire a webhook Emits an Automation Rule Triggered webhook event tagged with a trigger tag. See Firing a webhook.
Close case Closes an open case with the outcome Autoclosed. No later action or rule runs on the case. See Closing and archiving a case.
Archive case Archives an open case with the outcome Autoclosed. No later action or rule runs on the case.
Stop processing further rules No later rule runs on the case. The case does not change. See Stopping further rules.
Set severity / priority to Info, Low, Medium, High or Critical Sets the severity of an incident, or the priority of an investigation, to that level. See Setting the severity or priority.

Identifying a Slack user

For the Slack actions, enter an email address or a Slack member ID and click Resolve. DFIRe looks the user up through your Slack integration and confirms the match by showing their full name, @username, and member ID. Slack provides no lookup by username, so a plain handle on its own cannot be resolved.

Adding to the case team

Choose a role for each user you add. Lead, investigator, or viewer. Adding a user as lead makes them the case lead, replacing the current one. A rule can set only one lead, and a given user can appear in only one team action per rule (you cannot, for example, both add and remove the same person).

The rules of the case team apply. An investigator or a lead needs permission to edit evidence, and a role that can view all cases cannot be a viewer. The run history shows a user the rules do not allow as skipped, with the reason, and the rule's other actions still run.

Removing from the case team

Removal targets specific users you name in the rule, and removes them regardless of how they joined the case. Removing a team member also removes them from the case's Slack channel.

Adding and removing a group

The group actions use the role groups under Settings → Access Roles. Any role group works, including one that grants no permissions.

  • Add group to case team adds each active member as an investigator or a viewer. A member whose role cannot edit evidence joins as a viewer. A member who is already on the team keeps their role. A group cannot be the lead. To set a lead, add a user as lead in the same rule.
  • Remove group from case team removes every member of the group, deactivated accounts included. If the lead is a member, the case has no lead until another action sets one.

A rule reads the group's members when it runs. A user who joins the group later does not join the cases the rule has already run on. Each group can appear in only one group action per rule. If you delete a group, its action stays in the rule and fails on every run until you choose another group.

A case team can follow the case's project. Give each project a rule with the triggers Case opened and Project changed and a condition for that project. List its actions in this order: remove the groups of the other projects, add the project's group, then add the project's lead.

Setting the severity or priority

Each Set severity / priority action sets one level, and acts on an open case only. On a closed or archived case, and on a case that already has that level, the run history shows it as skipped. The timeline records the change with Automation as the actor, and the audit log records it with the rule's name.

A change a rule makes starts no other rule, so it does not fire the Severity / priority escalated and Severity / priority downgraded triggers. An incident's service level targets follow its severity only until someone acknowledges it. A later change does not move the targets.

Closing and archiving a case

Close case and Archive case act on an open case only. On a case that is already closed or archived, the run history shows them as skipped, and the run goes on. The rule's earlier actions run first, so a rule can assign the case and then close it, for example.

Both set the outcome to Autoclosed. DFIRe owns this outcome, so nobody can edit or delete it. People can choose it when they close a case by hand, for example to mark a case that a rule should have closed. Analytics leaves out cases with this outcome by default.

The closure has the effects of any closure. The timeline records it with Automation as the actor, and DFIRe records the closing time. Webhooks and email notifiers subscribed to Case Closed or Case Archived fire. A case that is still new is acknowledged when it closes.

The audit log records the change with Automation as the author. Nobody gets an in-app notification that the rule closed or archived the case, because the rule marks the case as one that needs no work. Notices from the rule's earlier actions, such as an assignment, still go out. The case then follows the retention policy like any other closed or archived case.

A case that a Case opened rule closes or archives gets no Slack channel and no Jira Epic. While an enabled Case opened rule has a Close case or Archive case action, a new case gets its channel and its automatic Jira Epic only after the rules have run. DFIRe then creates them only if the case is still open. A Slack action on a case without a channel shows as skipped in the run history. When a rule closes a case that already has a channel, DFIRe archives the channel. When a person closes a case, the channel stays open for follow-up work until the case is archived.

A closed case refuses new records. An integration that creates a case and then adds notes or evidence to it gets a 403 answer if a rule closed the case in between.

When a rule with Close case or Archive case has no conditions, the editor warns that the rule closes every case that reaches it. You can still save it.

Stopping further rules

Stop processing further rules ends the run without changing the case. Like every action, it runs only when its rule matches. Use it for exceptions. Place a rule with narrow conditions and this action above broader rules, and the broader rules do not run on the cases it matches.

Firing a webhook

The Fire a webhook action does not send a request itself. It emits an Automation Rule Triggered event carrying a trigger tag you choose. To deliver it, create an outgoing webhook subscribed to that event, and choose the same tag under Specific Automation triggers, or choose Any trigger. Delivery, signing, and retries are handled entirely by the webhook. A trigger tag takes letters, digits, dots, hyphens and underscores, up to 100 characters.

One webhook can serve both a runbook step and an automation rule. For example, tag both a runbook step and an automation with notify_ciso. Subscribe one webhook to both the runbook-step and automation events, and choose the tag for each, and either source pages the same destination.

See Webhooks for configuring the receiving webhook, its payload, and authentication.

Rule names and descriptions

The names and descriptions of drop rules and automation rules are plain text. DFIRe removes HTML markup from them and keeps the rest as typed. M&A noise and a < b stay as written, and <b>Noise</b> becomes Noise. Text in angle brackets, such as <[email protected]>, counts as markup, so DFIRe removes it too. A name must keep some text after this.

Run history

Each rule keeps a record of its recent runs, shown by expanding the rule. A run lists the triggering event, the case it ran on, and the outcome of each action, success, partial, or failed, with the reason for anything that did not complete. Slack invites and team changes are also recorded in the audit log, attributed to the system.

Running the rules on current cases

Drop rules, deduplication and automation rules act on new cases. To apply them to the open cases, for example after an alert storm, use Run automation rules on current cases under Settings → Automation.

The run takes the open cases only. To take only recent cases, enter a number of hours from 1 to 8760 in Cases created within the last. Leave it empty to take every open case. One run takes at most 1000 cases. When more open cases are in the window, DFIRe asks you to choose a shorter window.

DFIRe takes the cases oldest first. Each case goes through the same steps as a new case, with the rules as they are now:

  1. Drop rules. The first enabled drop rule that matches deletes the case, with its evidence items and notes. DFIRe cannot refuse a case that it already stores, so it deletes the case.
  2. Deduplication, when it is on. A case can repeat an earlier case that the run keeps open. DFIRe stores a duplicate record on the earlier case and deletes the repeat. When several earlier cases match, the most recent one receives the duplicate. The time window counts back from the repeat's creation.
  3. Automation rules with the trigger Case opened, in their order. Only the actions Close case, Archive case and Stop processing further rules run, and the other actions do not. The first matching rule with Close case or Archive case closes or archives the case with the Autoclosed outcome. A matching rule with Stop processing further rules ends the check, and the case stays as it is.

The run never deletes a case that holds files, custody records or indicators. It also leaves alone a case that someone created from a duplicate, while the duplicate record that links to the case exists. Such a case skips the drop rules and deduplication. The automation rules still apply to it, so the run can close or archive it. A later repeat in the run can still become its duplicate.

DFIRe knows a case created from a duplicate by that record. If someone deletes the record, or the case that holds it, the run treats the created case like any other case.

Preview and apply

Select Preview. The preview changes nothing. It shows how many cases the run would delete, add as a duplicate, close and archive, and how many it would leave as they are. The list shows each case that would change, oldest first, and what decides it: the drop rule, the earlier case or the automation rule. It also shows how many evidence items, files and notes each case holds.

Select Apply to make exactly the changes of the preview. The confirmation states the counts again. You cannot restore the cases that the run deletes, and the run deletes each repeat it adds as a duplicate.

While the run goes, the button counts the cases done. The run goes on to its end even if you close the page.

If the rules, the deduplication setting or the cases now give a different result than the preview, DFIRe changes nothing and asks you to preview again. A case created after the preview is not part of the run. The rules for new cases handle it.

DFIRe changes the cases one at a time. A case that it cannot change stays as it is, and the result lists it under Not changed with the reason. The other cases still change. The result shows how many cases changed.

The run sends no webhooks or notifications about cases it deletes, closes or archives. It leaves the match counts of drop rules and deduplication unchanged, because those count new cases. The audit log records each change and the whole run. You are the author of deletes and duplicate records. Automation is the author of closes and archives. DFIRe archives the Slack channel of each case it deletes, closes or archives.

Access control

Only a superuser views, creates, edits and deletes drop rules, the deduplication setting and automation rules, and runs the rules on current cases, in the web interface and through the API. The permissions a role holds do not change this. The audit log records each change with the superuser who made it as the author. Only readers of the whole audit log see those entries. See Audit log → Who sees what.

Drop rules and deduplication apply to every case, whoever creates it, superusers included.

A rule does not take on the permissions of its author. It runs on every case it matches, so a rule that adds or removes case-team members or groups changes case access on each of those cases.

The Slack actions can change case teams too. Auto-add investigators under Settings → Slack Integration is on by default. While it is on, a linked DFIRe user who joins the channel of an open case joins its team. A team member other than the lead who leaves the channel leaves the team. See Channel and team membership.