Your enterprise prospect's security team just asked for "your latest third-party penetration test." Your report is 64 pages long, full of CVSS vectors, request/response pairs, and a findings table that still shows two Highs as "open." Do you send the whole thing? A summary? A letter? And do you understand the report well enough to defend it on a call with their CISO?
This guide covers how to read a pentest report like a practitioner, how to verify and triage what's in it, and exactly what to send enterprise customers so deals keep moving without handing outsiders a map of your weaknesses.
For a growing SaaS company, the pentest report does two jobs. Internally, it is an engineering backlog with risk ratings attached. Externally, it is evidence. It shows procurement, vendor risk management, and auditors that an independent team tried to break your product and that you fixed what they found.
Those two audiences need very different documents. Engineers need every payload, endpoint, and reproduction step. A customer's third-party risk team needs to know who tested, what was in scope, when it happened, how the findings broke down by severity, and whether the serious issues are closed. Teams most often go wrong by mixing up the two, which either stalls deals or hands sensitive exploit detail to dozens of outside parties.
A well-built report from a reputable firm usually follows a predictable structure. Knowing where to look saves hours.
This is a one to three page narrative written for leadership. It should state the overall risk posture, the most significant attack paths, and the business impact in plain language. If you only have five minutes before a customer call, read this section and the findings summary table.
Watch for phrases like "an unauthenticated attacker could" versus "an authenticated low-privilege user could." That distinction changes the real-world risk a lot, and enterprise reviewers often ask about it.
Customers scrutinize this section more than any other, and internal teams skim it the most. Look for:
If your production application is not explicitly named in scope, a sharp reviewer will notice, and the report will not meet their requirement.
Each finding should include a title, severity, CVSS score and vector, affected assets, description, evidence, reproduction steps, impact, and remediation guidance. Strong reports also cite CWE identifiers and include a status field (Open, Remediated, Risk Accepted).
Appendices typically hold tool output, port scan results, lists of tested endpoints, and retest results. The retest appendix is especially useful in customer conversations because it proves the findings were closed.
A severity label is only a starting point. The CVSS vector string tells you why a finding got its score. You need that to explain the risk to a customer or to challenge an inflated rating.
Here is a quick way to decode any CVSS 3.1 vector from your report using the cvss Python library:
pip install cvss --break-system-packages
python3 - <<'EOF'
from cvss import CVSS3
vector = "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:N"
c = CVSS3(vector)
print("Base score: ", c.scores()[0])
print("Severity: ", c.severities()[0])
print("Clean vector: ", c.clean_vector())
labels = {
"AV": "Attack Vector", "AC": "Attack Complexity",
"PR": "Privileges Required", "UI": "User Interaction",
"S": "Scope", "C": "Confidentiality", "I": "Integrity", "A": "Availability",
}
for metric, value in c.metrics.items():
if metric in labels:
print(f"{labels[metric]:<22} {value}")
EOF
[cta]
In this example, AV:N (reachable over the network) and AC:L (low complexity) push the score up, while PR:L tells you the attacker needs a valid low-privilege account. In a multi-tenant SaaS product, that profile usually describes a broken object-level authorization issue, where any customer could read another customer's data. That is a High and deserves attention even though it requires a login.
When you disagree with a rating, don't argue over the base score. Apply environmental metrics instead (CVSS 3.1 modified metrics or the CVSS 4.0 environmental group). Document your compensating controls, then ask your testing provider to note the adjusted rating in the report.
Never patch blind. Reproducing a finding confirms you understand the root cause and gives you a reliable test for the fix.
In SaaS products, the findings enterprise buyers care about most are usually authorization flaws. Use two test accounts from different tenants and replay the request the tester documented:
# Tenant A token (attacker) requesting a Tenant B resource
export TOKEN_A="eyJhbGciOiJSUzI1NiIsInR5cCI6IkpXVCJ9..."
export TENANT_B_INVOICE="inv_8f2c91a7"
curl -s -o /dev/null -w "%{http_code}\n" \
-H "Authorization: Bearer ${TOKEN_A}" \
-H "Accept: application/json" \
"https://api.example.com/v2/invoices/${TENANT_B_INVOICE}"
# Expected after the fix: 403 or 404
# Vulnerable behavior: 200 with Tenant B data in the body
# Sweep a range of sequential object IDs with ffuf to confirm the fix holds
ffuf -u "https://api.example.com/v2/orders/FUZZ" \
-H "Authorization: Bearer ${TOKEN_A}" \
-w <(seq 100000 100500) \
-mc 200 -t 20 -rate 50
[cta]
In Burp Suite Professional, the Autorize extension automates this. It replays every request from a high-privilege session using a low-privilege session's cookie and flags any responses that match. If your report contains several authorization findings, Redfox Cybersecurity's API penetration testing engagements are designed around exactly this kind of cross-tenant, role-by-role testing.
External findings such as weak cipher suites, exposed admin panels, or outdated services are easy to verify and easy to reintroduce later:
# Full TLS assessment with machine-readable output
testssl.sh --severity MEDIUM --jsonfile-pretty tls_api.json https://api.example.com
# Pull only findings that are not OK
jq -r '.scanResult[0] | (.protocols + .ciphers + .vulnerabilities)[]
| select(.severity != "OK" and .severity != "INFO")
| "\(.severity)\t\(.id)\t\(.finding)"' tls_api.json
# Confirm a reported exposed service is now filtered
nmap -Pn -sV -p 22,3389,5432,6379,9200 --reason \
-oA verify_external 203.0.113.0/28
[cta]
If the original report flagged an exposed Redis or Elasticsearch port, the --reason flag in nmap shows whether a firewall now marks the port filtered or whether it still answers with syn-ack. Save the output, because it doubles as remediation evidence for the retest.
Internal network reports often describe a chain: a kerberoastable service account, a weak password, then a path to Domain Admin. Verify that the chain is actually broken:
# Check whether the service account still has an SPN set (kerberoastable)
impacket-GetUserSPNs corp.example.local/auditor:'S3cure!Pass' \
-dc-ip 10.10.0.10 -request -outputfile spn_hashes.txt
# Confirm SMB signing is now enforced on hosts flagged in the report
netexec smb 10.10.0.0/24 --gen-relay-list relay_targets.txt
# Re-collect AD data to confirm the attack path no longer exists
bloodhound-python -u auditor -p 'S3cure!Pass' -d corp.example.local \
-ns 10.10.0.10 -c All --zip
[cta]
An empty relay_targets.txt means SMB signing is enforced on every host scanned. Load the BloodHound zip and rerun the "Shortest Paths to Domain Admins" query. If the path described in the report still appears, the finding is not closed, whatever the ticket says.
Severity alone is a weak way to prioritize. Weigh it alongside exposure (internet-facing versus internal), data sensitivity (does the finding touch customer data?), and exploitability (is there a public exploit, or is it theoretical?).
Many providers now deliver findings as JSON or CSV in addition to PDF. Turn that export into a remediation tracker with an SLA date per finding:
# findings.json exported from the reporting platform
jq -r '
def sla(s): if s=="Critical" then 7 elif s=="High" then 30
elif s=="Medium" then 90 else 180 end;
["ID","Title","Severity","CVSS","Asset","Status","SLA_days"],
(.findings[]
| select(.status != "Remediated")
| [.id, .title, .severity, .cvss_score, (.assets | join(";")),
.status, sla(.severity)])
| @csv' findings.json > remediation_tracker.csv
# Quick severity distribution for the executive update
jq -r '.findings | group_by(.severity)
| map({severity: .[0].severity, count: length})[]
| "\(.severity): \(.count)"' findings.json
[cta]
Common SLA targets in enterprise contracts are 7 to 15 days for Critical, 30 days for High, 90 days for Medium, and best effort or 180 days for Low. Check your customer contracts and your SOC 2 or ISO 27001 vulnerability management policy, because auditors will measure you against whatever you committed to.
The retest turns your report from a list of problems into proof that you fixed them, so arrive prepared. Write each finding as a regression check so the fix survives future deployments. For web findings, a custom Nuclei template works well:
id: redfox-finding-07-debug-endpoint
info:
name: Exposed Spring Boot Actuator Heapdump (Pentest Finding 07)
author: appsec-team
severity: high
tags: regression,pentest,exposure
http:
- method: GET
path:
- "{{BaseURL}}/actuator/heapdump"
- "{{BaseURL}}/actuator/env"
matchers-condition: or
matchers:
- type: word
part: header
words:
- "application/octet-stream"
- type: word
part: body
words:
- "activeProfiles"
- "propertySources"
condition: and
[cta]
Run it in CI against staging on every release:
nuclei -t ./regression-templates/ -u https://staging.example.com \
-severity high,critical -jsonl -o regression_results.jsonl
# Fail the pipeline if any regression fires
test ! -s regression_results.jsonl || { echo "Pentest regression detected"; exit 1; }
[cta]
When your provider retests, ask them to update each finding's status in the report and issue a revised version or a retest letter. That updated document is what makes the customer conversation easy.
This is the question that matters most for revenue. In short, send the least sensitive document that meets the requirement, in a format designed for outside readers, under appropriate confidentiality.
An attestation letter (sometimes called a letter of engagement completion or a security assessment summary letter) is a one to two page document on the testing firm's letterhead. It typically includes:
For most mid-market deals, this letter alone gets you through the vendor security review. It reveals no exploit detail and can be shared broadly, including in a trust center.
Larger enterprises, financial services firms, and healthcare buyers often ask for more. Send the executive summary, the scope section, and the findings table with titles and severities. Strip out reproduction steps, payloads, internal hostnames, and screenshots. Share it only under NDA, ideally through a trust portal with view-only access and watermarking.
Reserve the full report for situations where a contract or regulator demands it, and share it only under NDA with a named recipient. Even then, it is reasonable to offer a screen-share walkthrough instead of handing over the file.
Never send a report with open Criticals or Highs and no remediation plan. Never send raw tool output, credentials captured during testing, or internal IP schemas. And never send a document whose file metadata exposes internal usernames or paths.
Sanitize every document before it leaves your organization:
# Inspect metadata on the outgoing PDF
exiftool Redfox_Pentest_Summary_2026.pdf | grep -Ei "author|creator|producer|title|subject"
# Strip all metadata and linearize for clean distribution
exiftool -all:all= -overwrite_original Redfox_Pentest_Summary_2026.pdf
qpdf --linearize --object-streams=generate \
Redfox_Pentest_Summary_2026.pdf Pentest_Summary_External.pdf
# Scan the extracted text for anything that should not leave the building
pdftotext Pentest_Summary_External.pdf - | grep -nEi \
"password|passwd|secret|token|10\.[0-9]+\.[0-9]+\.[0-9]+|192\.168\.|\.internal|\.local|BEGIN (RSA|OPENSSH)"
# Stamp a per-recipient watermark for traceability
qpdf --overlay watermark_acme_corp.pdf --repeat=1 -- \
Pentest_Summary_External.pdf Pentest_Summary_AcmeCorp.pdf
[cta]
Any hit from the grep step means the document needs another pass before it goes out. Per-recipient watermarks discourage forwarding, and if a copy ever surfaces, they tell you exactly whose it was.
Questionnaires such as SIG Lite, CAIQ, and custom enterprise spreadsheets ask nearly identical questions. Prepare standard answers once:
Consistent, confident answers backed by a current attestation letter can shorten security reviews from weeks to days. A last test that is more than a year old, or one that skipped your API, will surface here first. If either applies, schedule a fresh web application and API penetration test with Redfox Cybersecurity before your next big renewal or RFP.
Budget questions come up fast once customers start asking for reports. As a reference point, Redfox Cybersecurity publishes starting prices on its site and notes that final pricing depends on scoping:
The single-report bundle is especially useful for enterprise sales because it gives you one attestation letter covering everything a vendor risk team asks about. For a SaaS company selling into regulated industries, the Redfox Cybersecurity cloud configuration review pairs naturally with application testing, since buyers increasingly ask about IAM and storage exposure alongside application security.
As for frequency, most enterprise contracts and compliance frameworks expect at least annual testing. Add a targeted test after major architecture changes, such as a new authentication system, a migration to a new cloud provider, or the launch of a public API.
A pentest report only helps you if you can understand it, act on it, and present it credibly. If you want a report built for both your engineers and your enterprise customers, with clear scoping, retest letters, and a shareable attestation, talk to Redfox Cybersecurity about your next engagement.