We find what auditors will find — before they do. CI/CD security assessments with AI-verified fixes, board-ready evidence, and an air-gapped Sovereign option where your data never leaves your perimeter. Every fix is checked against your workflow — we tell you honestly which are ready to paste and which need your judgment.
Ready to start? Email us — we reply within 24 hours on business days.
From $100/developer · One-time fee · No subscription
A founding client isn't a beta tester — you receive the identical deliverable set a paying client gets, and in exchange your written feedback becomes part of how the practice earns the right to charge.
The complete assessment, free: executive summary, remediation guide with per-fix code (paste-ready or honestly labeled needs-judgment), SARIF, SBOM, compliance mapping, and the verification record. Same depth a paying client receives — nothing is held back.
15 minutes of an engineer's honest feedback, in writing: does the assessment hold up? What's missing? What would you pay for? No calls required. A short email reply is perfect.
The analysis is built and verified by AI agents under deterministic checks, and we're transparent about that. Before we charge anyone, we want a handful of strong engineering teams to tell us — in writing — whether the work holds up. Your feedback shapes the product.
Limited cohort: founding terms apply while we validate the practice. Standard pricing below applies afterwards — the assessments stay identical.
Automated tools find vulnerabilities. They don't tell you which ones matter, how to fix them, or how to prove to an auditor that you did.
74.5% of CI/CD findings in GitHub's top 50,000 repos trace to unpinned actions with mutable tags. A single compromised tag gives attackers code execution in your production pipeline.
Hardcoded tokens, leaked credentials, and overprivileged service accounts — present in nearly 9 out of 10 CI/CD assessments we've run.
SOC 2, ISO 27001, and NIST CSF auditors now ask about CI/CD pipeline security. If you can't evidence it, you fail — or pay for remediation.
From a one-time deep assessment to sovereign analysis that never touches the cloud. Choose what fits your risk profile.
Full CI/CD security assessment with CVSS 4.0 severity scoring, OWASP CI/CD Security Top 10 classification, business impact analysis, and AI-verified fixes with honest per-fix status. Includes SOC 2, ISO 27001, and NIST CSF audit evidence.
Zero data egress. We run the Reckon Intelligence Stack on isolated infrastructure — your pipeline metadata never leaves your perimeter. Built for defence, fintech, and regulated industries.
Code changes. Actions change. A one-time audit is stale in 3 months. Quarterly refreshes keep your compliance evidence current and your pipeline secure.
Need it faster? Express turnaround — 1 business day for +20% surcharge. Available on any tier. Deposit and access confirmation required by 10:00 GMT.
Turnaround times are indicative for standard-complexity repos. After the initial scan, we confirm your final delivery date based on the assessed finding count — repos with 150+ findings may need additional time.
We don't sell tools. We sell outcomes — audit evidence, proof of diligence, and CI/CD pipelines you can defend to a board. The practice is led by a Founder & Operations Lead and built on AI agents under deterministic verification — no security-team payroll is priced into the work.
Enable on any engagement: three independent verification passes with consensus required before a finding is confirmed. Contested findings get a 4th-pass tiebreaker. Either way, deterministic fact-checking validates every finding against your source files — on every engagement, no exceptions.
Our scanner patterns are continuously updated from the CISA Known Exploited Vulnerabilities catalog and GitHub Security Advisories. Newly disclosed CVEs are continuously evaluated for CI/CD scannability and validated before joining our detection database.
Every finding is scored using CVSS 4.0 with full vector strings, contextualised for business impact, and classified under the OWASP CI/CD Security Top 10. We tell you what matters and why — not a prioritised YAML dump.
Our reports are written for auditors and executives, not engineers. SOC 2, ISO 27001, and NIST CSF evidence included.
For regulated industries, we run on isolated infrastructure. Your pipeline logic and secrets never leave your perimeter.
Deep Clean delivered in 3 business days (5 for 26+ devs), confirmed after the initial scan. Not weeks. Not a retainer. A one-time engagement with a clear deliverable.
$100/developer one-time. Subscription tools charge $150-$200+/dev/year — every year. We are roughly 50% cheaper in year one — and you own the report forever.
For pentest firms and security consultancies: add CI/CD assessments to your offering. We do the work, you keep the client relationship.
We assess the CI/CD pipeline configuration attack surface. Every finding is classified under the OWASP CI/CD Security Top 10. We tell you what's in scope and what isn't, up front.
Org-level controls require API access and are outside static pipeline analysis. The OWASP CI/CD Security Top 10 covers all 10 categories — we have strong depth in 7, and disclose the 3 that require org-level access.
Founding clients pay $0 — see the founding-client program above. The prices below are the standard rates founding terms convert to afterwards.
No subscriptions. No seat licences. No hidden fees. Pay once, own the report, use it for your audit.
| Service | Per Developer | Min / Max | Notes |
|---|---|---|---|
| Deep Clean | $100 | $500 - $2,500 | One-time · Full assessment · 7-day support |
| Sovereign Shield | $150 | $750 - $3,750 | One-time · Air-gapped · Zero data egress |
| Express (any tier) | +20% | — | 1 business day turnaround · Same deliverables · Requires deposit by 10:00 GMT · Standard-complexity repos |
| Quarterly Refresh | $25 (Sovereign: $40) | $125 min | Requires prior Deep Clean or Sovereign · Billed quarterly |
| Re-scan (after fix) | $50 | -- | Discounted follow-up scan |
| Extended Support | $250 | Flat | +7 days of email support |
| Monthly Retainer | $5,000 | Flat | Weekly threat briefings + on-demand re-scans |
| CI/CD Add-on (Partnership) | $1,000 | Flat | For pentest firms · You keep $200 |
Volume discount: 10% off for assessments of 3+ repositories booked simultaneously.
All engagements subject to 50% deposit to begin. Balance due before full report delivery. Ask about our payment terms for enterprise engagements.
All prices in USD. Developer count = unique human commit authors in the trailing 12 months (dormant repos: all-time). Minimum 5 developers. Enterprise teams (50+) — contact us for custom pricing. Partnership pricing available for pentest firms and security consultancies.
Every finding ships with evidence, plain-language context, a grounded fix, and an honest per-fix status — the same three the deliverable reports on every assessment: ✅ Paste-Ready (verified against your workflow, safe to apply), ⚠️ Needs Judgment (correct guidance that depends on your context), or ℹ️ Advisory (hardening guidance where no code change is required). Illustrative example — not from any client repository.
uses: actions/checkout@<vN>💰 Business Impact: A compromised tag injects attacker code into your build with your existing permissions. Pinning to a commit SHA makes the reference immutable.
What this means: A mutable tag can be moved by whoever controls the action repo — or by an attacker who compromises it. Your pipeline executes that code on every build.
# BEFORE (vulnerable):
uses: actions/checkout@<vN>
# AFTER (verified SHA):
uses: actions/checkout@0123456789abcdef0123456789abcdef01234567 # was @<vN>
on: push / merge_group — no concurrency: block💰 Business Impact: A cancelled-in-progress deploy can stop mid-release; races can ship a broken build over a good one. Grouping keeps one run per ref.
What this means: Simultaneous runs can race each other — two deploys interleaving, stale artifacts overwriting fresh ones, duplicate cloud resource churn.
# Guidance — add a concurrency block, then decide:
concurrency:
group: deploy-${{ github.ref }}
# cancel-in-progress: true # ONLY for disposable CI runs.
# For deploy workflows releasing to production, leave it OFF:
# cancelling an in-flight release mid-publish is worse than queuing.
# Choose the group key to match how you want runs serialized.
This fix needs your call: the safe setting differs for CI vs deploy workflows. We flag it honestly instead of guessing.
No workflow step produces a CycloneDX / SPDX SBOM💰 Business Impact: When the next log4shell-scale CVE drops, answering "are we exposed?" takes days of manual dependency hunting instead of one query against an inventory.
What this means: An SBOM is a machine-readable inventory of every component your build ships. Enterprise buyers and auditors increasingly ask for one outright.
# Guidance — nothing is broken; this is a hardening recommendation.
# Add an SBOM step once per build. Tooling depends on your stack, e.g.:
# - name: Generate SBOM
# uses: <sbom-action>@<pinned-commit-sha>
# Publish the output next to your release artifacts.
Advisory: no fix is forced. We recommend, you pick the tooling — no universal paste invented for show.
Every deliverable includes: full workflow map with severity + OWASP CI/CD classification, per-fix status for every finding, copy buttons, SARIF + SBOM exports, compliance mapping, and an executive summary for auditors.
Let's have that conversation before it becomes a finding.