n8n-nodes-halopsacomplete
The AI review rates the findings as likely false positive, but the risk score (53/100) still counts them. The review covers version 5.5.4, not the analysed version 5.5.5.
Analysis record
- Analysed
- Today 22 November 2026
- Version
- v5.5.5
- Artifact
- SHA256 15656e21…601f64
- Detail
- Partial list
Is n8n-nodes-halopsacomplete safe?
This package adds a HaloPSA node to n8n, so a workflow can read tickets, surveys and asset records from a HaloPSA helpdesk and act on them. It declares no host permissions of its own. HaloPSA credentials come from n8n's credential store, and the node calls whatever instance URL you configure there.
The network list needs reading before it alarms anyone. Entries such as updatefields.site, responsedata.tickets and body.site are chained JavaScript field lookups inside the compiled action files, and the scanner split each one at the dot. The one real network finding, NET-FETCH-dist/nodes/HaloPSA/actions/surveys/getAll/description.js, sits in a file that builds the request description for the surveys list action in the editor.
Nothing reads .ssh, .aws, .kube or browser password stores, and no secrets turn up in the code. Tool descriptions carry no hidden instructions for an AI agent, which matters for n8n because agent nodes can pass those descriptions to a model.
The scanner counted thousands of hits because dist/nodes/HaloPSA/ ships minified build output. Code-style rules fire on any minified JavaScript, and the indicator list fills with dot-split property names. Read the underlying files and the picture is a plain API wrapper.
Evidence ledger
Ranked by severity · findings with a source location link to the code viewer
YARA Rule Matches
9 rules| Severity | Rule | Hits | Files | Metadata |
|---|---|---|---|---|
| LOW | postinstall file download | 13 | dist/nodes/HaloPSA/actions/tickets/getAll/description.jsdist/nodes/HaloPSA/pagination.d.tsdist/nodes/HaloPSA/actions/raynet/getAll/description.js +10 more | - |
| LOW | credential env files | 3 | dist/package.jsonREADME.mdpackage.json | - |
| LOW | postinstall persistence mechanism | 4 | dist/nodes/HaloPSA/actions/invoices/getAll/description.jsdist/tsconfig.tsbuildinfodist/nodes/HaloPSA/actions/invoices/create/description.js +1 more | - |
| LOW | postinstall file manipulation | 495 | dist/nodes/HaloPSA/actions/attachments/delete/description.jsdist/nodes/HaloPSA/actions/webhookEvents/delete/description.js.mapdist/nodes/HaloPSA/actions/releaseNoteGroups/delete/index.js +492 more | - |
| LOW | postinstall system command | 435 | dist/nodes/HaloPSA/actions/actions/getById/index.jsdist/nodes/HaloPSA/actions/sites/update/execute.jsdist/nodes/HaloPSA/actions/purchaseOrders/update/execute.js +432 more | - |
| LOW | postinstall crypto operations | 15 | dist/tsconfig.tsbuildinfodist/nodes/HaloPSA/actions/securityChecks/getOldEncryption/index.d.tsdist/nodes/HaloPSA/actions/securityChecks/index.d.ts +12 more | - |
| LOW | postinstall network communication | 25 | dist/nodes/HaloPSA/actions/users/update/description.jsdist/nodes/HaloPSA/actions/interfaces/sites.d.tsdist/nodes/HaloPSA/actions/ticketTypes/update/description.js +22 more | - |
| LOW | postinstall obfuscation | 8 | dist/nodes/HaloPSA/actions/fieldInfo/getAll/description.jsdist/nodes/HaloPSA/actions/assets/update/description.jsdocs/attachments.md +5 more | - |
| LOW | postinstall environment access | 1 | dist/nodes/HaloPSA/actions/interfaces/actions.d.ts | - |
Network Indicators
Concrete URLs, domains, IPs, emails, and hashes extracted from the analyzed artifact.
Loading network indicators
Long lists can take a few seconds.
Finding Categories
YARA Rules Matched
9 rules(999 hits)AI Security Report
AI Security Review
Evidence context: threat category none; evidence quality moderate.
n8n-nodes-halopsacomplete 5.5.4 is a community node for n8n that talks to the HaloPSA helpdesk API. Everything shipped is compiled output under dist/nodes/HaloPSA/, built by a CI account named GitHub Actions. A node like this is ordinary glue code: action files that assemble request descriptions, plus transport code that calls a base URL taken from n8n credentials.
Tool poisoning gets zero matches across the scan, and the file structure supports that. The description.js files hold field labels and endpoint paths for n8n's editor forms. There is no directive addressed to a model, no invisible character run, and no attempt to hide a tool call inside metadata. n8n can hand tool descriptions to agent sessions, so a poisoned node would matter here. None of that pattern shows up.
Credential handling is equally quiet. No finding touches .ssh, .aws/credentials, .kube/config or application_default_credentials.json, and the secret category is empty. The node declares no host permissions. HaloPSA API keys arrive through n8n's credential store at runtime, which is how every node in that store works. Nothing in the code harvests environment variables and nothing posts them to a third-party host.
The only network finding is NET-FETCH-dist/nodes/HaloPSA/actions/surveys/getAll/description.js, which lands in the description file for the surveys list action. That file assembles request metadata for the editor. It does not send data anywhere by itself.
The endpoint list looks alarming until you read the strings: updatefields.site, body.site, responsedata.tickets, ticketdata.site, qs.group. These are property accesses inside compiled JavaScript, and the extractor split each chain at the dot and kept the tail. email.com and 4c.gg come from similar fragments. None of them correspond to a service this package contacts.
The strongest counterargument is volume. Roughly 3,500 hits on a package published through an anonymous-looking CI account, in a store with light vetting, deserves a second look. The volume collapses under inspection. Malware signatures are zero, obfuscation is zero, and the code-style rules fire on minified field-mapping code that every n8n community node carries. What remains is a wrapper around a documented API.
Key Reasons
- Zero tool-poisoning matches; description.js files contain editor field labels and endpoint paths, with no hidden directives aimed at a model
- No credential-path access: no findings touch .ssh, .aws/credentials, .kube/config or application_default_credentials.json, and the secret category is empty
- The single NET-FETCH sits in dist/nodes/HaloPSA/actions/surveys/getAll/description.js, a request-description file for the editor
- Endpoint list entries such as updatefields.site, body.site and responsedata.tickets are dot-split property chains from compiled code, not contacted hosts
- Malware signature, obfuscation, dependency and secret categories are all empty; the volume is bundled minified build output
False Positive Considerations
- IoC extractor splitting JavaScript property chains into fake hostnames
- YARA code-smell rules firing on minified bundled output in dist/
- No tool-poisoning or credential findings to review
Reviewed 2026-10-27 · version 5.5.4; recommended action: suppress false positive; model confidence 83%.
n8n version history
Risk trend by version
6 analyzed versions. Each point is the latest successful scan for that version; failed zero-score scans are hidden. Dates: First seen by Risky Plugins.
Pick any point on the chart, then open the Source tab to read that version's code.
Source Code Not Available
Source code is not available for this version of the extension.