After you submit
After you submit, each task is held to the intake standard, then audited in an isolated sandbox. It comes back clean, or flawed with its proof.
After you submit, each task is held to the intake standard and then audited in an isolated sandbox. It comes back clean, for the exact files that were audited, or flawed, with the flaw and its proof. A task that cannot be audited comes back with a request for what is missing.
The three steps
| Step | What happens | What you see |
|---|---|---|
| 1 | Your tasks are held to the intake standard. Nothing runs until you confirm the list. | "Reading your source", then the preview |
| 2 | A task that misses a rule comes back with one line for each thing that is missing. You supply it. | "Checking", then "Needs from you" |
| 3 | Each task is audited in an isolated sandbox. It comes back clean, or flawed with its proof. | "Running", then "Clean" or "Flawed" |
Your tasks are held to the intake standard
You send one archive of task folders, with a collection name. Horizon Audit finds every task in the archive and pins each one to a digest. While the archive is read, the submission shows "Reading your source". You then see a preview and choose which tasks go on. Nothing runs until you confirm the list. Submit tasks has the details.
After you confirm, each task shows "Checking". It is held to the intake standard, rules IS-1 to IS-9. The task screen shows a checklist of the rules.
A task that misses a rule comes back to you
A task that misses a rule shows "Needs from you", with one line for each thing that is missing. Each line names the rule, the file, and what to supply. No line says how to write it.
"Needs from you" is not a verdict. The task is not judged, and it is never counted as clean. Fix and resubmit says how a request is met.
Each task is audited
A task that meets the intake standard is audited in an isolated sandbox. The task is run as it is written. It is never edited to make it run.
On the way to its result, a task shows these labels:
| Label | What it means |
|---|---|
| Scoring | The task is being given a difficulty band. |
| Queued | The task waits for a sandbox. |
| Running | The task is being audited. |
| In review | An engineer is looking at the task. |
| Clean | The verdict is clean. |
| Flawed | The verdict is flawed, and the finding is published. |
A band is one of Basic, Standard, Complex or Specialist. A band says how much work an audit of the task is. Task states lists every label.
Tasks do not wait for each other
Each task moves on its own. A result appears as soon as its task is done. One task that waits for you does not hold up the others.
Who stands behind the verdict
The verdict is computed from the saved output of runs. An engineer confirms every flawed verdict before you see it. You never see a draft.
What comes back
A clean task shows "Clean". The verdict holds for the exact files that were audited, because each task is pinned to a digest of its files. A flawed task shows "Flawed" and has a finding with its proof. Read a verdict says what each result contains.
A changed task is sent again
A changed task is sent again, in a new archive under a new collection name. It is audited from the start, and you withdraw the old task. A verdict holds only for the files that were audited, so it does not cover the changed task. Horizon Audit does not suggest a fix and never edits your task.