Evidence and Artifacts
What scan artifacts exist and how reviewers should use them.
What this is
Artifacts are local files that preserve scanner evidence, graph paths, decisions, and optional model analysis.
When to use it
Use artifacts when reviewing a finding, exporting proof, or debugging a rule.
Command or example
ls .agent-permit/runs/<run_id>Output to expect
Important files include:
summary.mdrisk-report.mdpermit.yamlraw-findings.jsongraph-paths.jsoncontrols.jsonrun-metrics.jsonagent-investigation.md
How to interpret it
Use Markdown files for human review. Use JSON and YAML files for automation, dashboards, tests, and AI analysis.
Common mistakes
- Publishing generated artifacts without redaction.
- Reading
agent-investigation.mdwithout checking citations. - Moving artifacts out of the run folder and breaking path context.