Leadership asks for “last week’s security report,” you paste a screenshot of the live Findings list—and next week the same link shows different numbers auditors will not accept. Usually the scan ran; you treated a live issue list as a report.
Takeaway: A report is a frozen snapshot: severity distribution, gate result, scope, and engine versions land in the Document and do not auto-rewrite on the next check. Day-to-day work uses alerts and Findings; reviews, monthly packs, and external delivery use reports. Job-scoped reports require completed jobs. Formats follow the console: commonly HTML, JSON, CSV, PDF; finding details may also download as SARIF.
How to read severity and paths: see Reference. Below is how to export (audience, scope, monthly, formats, share). Filtering by SAST/deps/image/secret/license: see capability-scoped reports.
1. Report vs Finding vs alert
Engine output → Finding (items to fix)
→ Alert (high-signal ops)
→ Report (downloadable snapshot frozen at generation)
| Object | Changes by itself? | Typical use |
|---|---|---|
| Finding / alert | Yes—new checks, suppressions, fixes | Daily triage |
| Report | Frozen after generation | Review, archive, external |
Upload success or registry ingest ≠ scanned ≠ gate passed ≠ a deliverable report exists.
2. Four audiences (pick a template)
Console create wizard commonly shows four audience cards (UI labels win):
| Audience | Reader takes away | Default feel |
|---|---|---|
| Management posture | How bad, vs last period, what to decide | Short, conclusion-first; often PDF; avoid dumping finding details |
| Engineering remediation | What to fix first, paths, CVE order | Full detail; often HTML/CSV; can enable findings + SARIF |
| Third-party / external | What was tested, method/versions, who approved exceptions, verifiability | Stronger redaction; often PDF+JSON; may attach evidence pack |
| Custom | Free chapter mix | Team templates; still under redaction and RBAC |
Legacy enums like EXECUTIVE / ENGINEERING / EXTERNAL map to management / engineering / external—use console copy on materials; do not print raw enums externally.
3. Shortest console path (text only)
No screenshots needed—follow wizard labels:
- Sidebar Reports → Create (or Quick start: monthly posture / engineering checklist / job-scoped report).
- Audience: switching may reset chapters/display while keeping scope/filters (follow on-page hints).
-
Scope one of:
- All assets current state (still freezes as a snapshot);
- Time range (last month / last 30 days—monthly path);
- Selected jobs: only completed; never use failed/incomplete as release evidence.
- Filters (optional): severity, engine, capability. For “SAST-only” etc., see capability article.
- Detail/chapters: management stays brief; engineering opens finding details if you need SARIF; job-scoped mode may disable trend chapters—follow hints.
- Formats: at least one; primary format sorts first in download. CSV may be a ZIP of chapter CSVs.
- Generate → Generating → Ready (READY) before download. Font failures (e.g. missing CJK for PDF) mark the report failed—do not treat partials as success.
- External: header often Unpublished; Publish then share. Shares may set TTL and password; access should be audited. Create/download: analyst or admin (RBAC).
To narrow a “vuln scan report” by severity: raise minimum severity or keep only CRITICAL/HIGH so INFO noise stays out of management PDFs.
4. Monthly security report
| Step | Do this |
|---|---|
| Entry | Quick start “monthly posture” or Create → management audience |
| Scope | Time range → last month (or last 30 days) |
| Primary format | PDF for minutes/archive |
| Questions to answer | Risk up/down, asset ranking, gate coverage, exception drift |
Weekly reports reuse the same template with a shorter window. Scheduled subscriptions, if offered, should use relative windows; otherwise generate manually. Mail/channel config follows the console.
What auditors need beyond a monthly PDF: see compliance evidence article—do not call a monthly report an assessment report.
5. Export formats
| Format | Use | Watch-outs |
|---|---|---|
| HTML | Browser reading, internal share | Self-contained preview per page |
| Management / formal external | Missing CJK fonts can fail generation | |
| CSV | Spreadsheets | Often a ZIP of chapter CSVs |
| JSON | Integration / archive | Full Document; external JSON should redact repo URLs / member names |
| SARIF | Findings into IDE / tickets / GitHub | Findings only; unavailable without finding-details chapter; management/external defaults often lack details; quality-only or license-only scopes often lack SARIF—trust the download menu |
Multi-format downloads from one Document should match numerically. External evidence packs (PDF/HTML/JSON + method/coverage, often with checksums) may attach by default—console wins.
6. Share and external delivery
- Shares: TTL, password, audited access.
- External: redact paths, repo URLs, member IDs; never put secret bodies or recoverable snippets in external files (engineering audience for locations; external for counts).
- Unpublished external reports should block share.
- Revoke old shares after republishing so auditors do not get drafts.
7. Failure table
| Symptom | Cause | Fix |
|---|---|---|
| Download 4xx / not ready | Still generating | Wait for READY |
| Numbers ≠ home page | Comparing live Findings to an old report | Snapshot semantics; regenerate for new numbers |
| Job-scoped empty | Job incomplete/failed | Select completed jobs only |
| No SARIF menu | Finding details off, or scope without finding-like items | Engineering audience + details, or CSV/JSON |
| PDF failed / tofu glyphs | Fonts missing | Install fonts per ops runbook; do not deliver failed state |
| External PDF has repo URLs | Wrong audience/redaction | External template + publish, then share |
| Pentest template pasted in | Product does not do pentest/DAST reports | Stay in AppSec scan report scope |
FAQ
If another scan runs after generation, does the report update?
No. Generate a new report for new results.
Host / Qualys-style asset reports?
No. Scope is code and artifact supply-chain scanning, not host inventory.
Does finishing a scan mean MLPS pass?
No. Reports are one evidence input; assessors and process decide. See compliance evidence article.
Not allowed to download?
Create/download usually need analyst or admin; follow tenant RBAC.
How does this relate to capability-scoped reports?
This piece covers freeze and delivery (audience, window, formats, share). Filter recipes (SAST-only, image job-only, …) are in the capability-scoped reports article. Learn export first, then narrow.
Summary
- Report = frozen snapshot; Finding = daily triage.
- Audience → scope → completed jobs → READY download.
- Monthly = management audience + time window; SARIF is findings-only.
- External: publish then share; no raw secrets outbound.
Reference articles
- How to read scan reports: https://www.anguskit.com/en/blog/how-to-read-security-scan-reports
- Capability-scoped reports: https://www.anguskit.com/en/blog/capability-security-reports-sast-sca-secret-license
- Audit evidence checklist: https://www.anguskit.com/en/blog/compliance-audit-security-scan-evidence
- Gate thresholds: https://www.anguskit.com/en/blog/angussecurity-scan-policy-severity-thresholds
- Replace scanner scripts: https://www.anguskit.com/en/blog/replace-semgrep-trivy-gitleaks-with-one-scan
- Ask about highs, findings, and gates via MCP: https://www.anguskit.com/en/blog/angussecurity-mcp-server-design
- Docs: https://www.anguskit.com/en/docs/security
Top comments (0)