Skip to content
Closed
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 17 additions & 0 deletions docs/src/content/docs/reference/safe-outputs.md
Original file line number Diff line number Diff line change
Expand Up @@ -1381,6 +1381,23 @@ safe-outputs:

**Variables**: `{workflow_name}`, `{run_url}`, `{triggering_number}`, `{workflow_source}`, `{workflow_source_url}`, `{event_type}`, `{status}`, `{operation}`

### Security Metadata (`secrecy`, `integrity`)

Any safe output tool call can optionally include `secrecy` and `integrity` metadata fields. Unlike frontmatter options, these are set in the **agent's output** (the JSON the agent sends when calling a safe output tool), not in the workflow's `safe-outputs:` configuration.

- **`secrecy`**: Confidentiality level of the message content. Values: `"public"`, `"internal"`, `"private"`.
- **`integrity`**: Trustworthiness of the data source. Values: `"low"`, `"medium"`, `"high"`.

When present, both values are displayed in the workflow step summary.

**Example — agent output including security metadata:**

```json
{"type": "create_issue", "title": "Security report", "body": "...", "secrecy": "internal", "integrity": "medium"}
```

See [Secrecy](/gh-aw/reference/glossary/#secrecy) and [Integrity](/gh-aw/reference/glossary/#integrity) in the Glossary for definitions.

## Related Documentation

- [Threat Detection Guide](/gh-aw/reference/threat-detection/) - Complete threat detection documentation and examples
Expand Down