Status: working playbook, alpha project. habitable is an independent, unfunded, bus-factor-1 FOSS project. The alpha caveat in the README and threat model stays until, among other things, an independent security + cryptographic audit is done. This document is how we go get that audit paid for. It covers all three paths — grants, pro-bono, and paid — and ends with reusable application and RFP drafts.
Scope. This file covers recruiting/funding the security & cryptographic auditor (recruitment role 1) and, where a channel bundles it, the accessibility tester (role 2). The housing/tenant lawyer (role 3) and the pilot partner (role 4) are handled in separate recruitment docs and are out of scope here.
Research provenance. The program facts below come from a funding-research pass dated 2026-06-17. Anything that could move between then and when you actually hit send is tagged
[RE-VERIFY]. Treat those as "check the live page first." See §7 Re-verification checklist.
- Pursue the two channels actually built for us, in parallel, first: OTF Security Lab (free, third-party audit they pay for) and OSTIF (free facilitation + fundraising for the audit). Both are designed for an unfunded, single-maintainer, crypto-heavy FOSS tool.
- Add GitHub Secure Open Source Fund as a strong, lower-friction parallel bet — our existing security posture is exactly what it rewards — accepting that it is a cohort/uplift program whose money goes toward an audit rather than buying one outright.
- Treat NLnet, Sovereign Tech, and Mozilla MOSS as "monitor / indirect," not "apply today" — eligibility mismatches are real and documented below.
- Keep a costed paid-firm RFP ready so that if a funder (or OSTIF) says "yes, we'll fund it," we can scope and award within days, not months.
- Re-verify every program fact tagged
[RE-VERIFY]before sending.
The single asset that makes every one of these cheaper is already in the repo:
the frozen, hash-pinned threat-model baseline B1
(../audits/threat-model-baseline.md) plus the
reviewer onboarding doc (../audits/onboarding.md).
Lead with those everywhere.
| Dimension | Grant / facilitated (OTF, OSTIF, GitHub SOS, NLnet…) | Pro-bono (uni clinics, OWASP, academic labs) | Paid firm (Trail of Bits, Cure53, ROS, Quarkslab…) |
|---|---|---|---|
| Cash out of our pocket | $0 (they pay the auditor or pay us) | $0 | ~$30k–$200k for a real crypto+app audit (OSTIF's cited range) [RE-VERIFY pricing] |
| Our effort to land it | Medium–high: application, interview/video, scoping calls; eligibility framing matters | Low–medium per contact, but high volume of cold outreach; opportunistic | Low to get a quote; the cost is money, not effort, once funded |
| Time application → report | Long: ~3–6+ months end to end (intake + scope + auditor assignment + audit window) | Variable / unpredictable; can be a semester (clinics) or indefinite (labs) | Fast once funded: weeks to a few months for the engagement itself |
| Certainty of landing it | Medium. OTF's "internet-freedom" lens and OSTIF's "critical/widely-used" bias are the chief risks for a small alpha tool | Low–medium. Reachable, but depth on advanced primitives is the gap | High if funded — it's a contract |
| Depth / credibility of result | High (OTF/OSTIF route to top firms) to high-uplift (GitHub SOS is education+funding) | Usually shallow on crypto: org-level / vuln-assessment / web-surface, not Ed25519/X25519/RFC-3161 primitive review | Highest and most citable; publishable report |
| Publishability | Usually yes (OTF/OSTIF publish reports) | Sometimes (labs publish; clinics often don't) | Yes — that's the point |
| Best use for habitable | Primary path. Get the deep crypto+app audit paid for | Complement / warm-up. Web-app & packet.html surface; a second set of eyes before the deep review |
The thing the grant pays for. Keep an RFP ready to execute on funding |
Strategy that falls out of the table: run grant + pro-bono in parallel
(they don't conflict and the pro-bono breadth can de-risk the paid scope), and keep
the paid RFP costed and ready so a "yes, funded" converts immediately. Reserve
the deep cryptographic review for an OTF/OSTIF-facilitated or paid specialist firm;
use pro-bono channels for breadth and the web-app/packet.html attack surface.
Programs are ordered by realistic fit for habitable today, best first. Each has a one-line verdict, the concrete steps, what to attach, and the honest fit risk.
Verdict: Best-targeted free audit channel. They contract and pay an independent auditor; we receive the audit, not cash. Projects that are not otherwise OTF-funded may apply specifically for an audit.
- Open the project-side audit application at
apply.opentech.fund/security-lab/[RE-VERIFY URL + that project-side intake is open] - Lead with the threat model (
../threat-model.md) and baseline B1 (../audits/threat-model-baseline.md) - State plainly that the alpha caveat is gated on exactly this review (a precondition on the v1.0 gate)
- Lean hard into the at-risk-user + privacy-claim-validation angle: tenants documenting habitability problems who face landlord retaliation; the tool's job is tamper-evidence under an adversary. This is the framing that answers OTF's "internet freedom" lens.
- List the specific stated claims to validate (ChaCha20-Poly1305 at rest, scrypt-wrapped key, Ed25519 signing, X25519 sealed-box sync, RFC 3161 timestamping + archive re-timestamping, SHA-256 fixity, hash-linked custody chain, standalone verifier)
- Point to synthetic-data-only evaluation via
habitable demo(no real tenant data, ever) and the onboarding doc - Note expected turnaround and do not confuse this with the March 2026 RFP that recruited auditors, not projects
Fit risk (state it, don't hide it): OTF frames around internet-freedom /
repressive-environment users. A domestic US tenant tool is a defensible but not
slam-dunk fit. Mitigation = the at-risk-user + retaliation + privacy-tool framing
above. Timeline: rolling intake; plan 3–6+ months end to end. Confidence
in the facts: high that the project-side form exists; the eligibility lens is the
real uncertainty. [RE-VERIFY queue depth / turnaround]
Verdict: The natural facilitator for a bus-factor-1 maintainer who can't run an RFP alone. They scope, pick a reputable firm (Quarkslab, Shielder, Ada Logics, …), run the engagement, publish the report — and help find/split the funding (Linux Foundation/OpenSSF, Sovereign Tech Bug Resilience, CNCF, ~10 funders).
- Do their pre-audit self-checks (best-practices guide) first — we already pass much of it:
make verifygate, CodeQL, pip-audit, Dependabot, SHA-pinned CI, SBOM+provenance releases - Read the "Get an Audit" page:
ostif.org/get-an-audit/[RE-VERIFY page + intake] - Email
contactus@ostif.org(use the sponsorship/info@contact for the funding conversation)[RE-VERIFY addresses] - Be upfront about our situation: unfunded, bus-factor-1, alpha — and ask directly whether they can slot habitable into a funded program (e.g. Sovereign Tech Bug Resilience Program, which funds audits through OSTIF) or help raise/split the cost
- Attach baseline B1 and the onboarding doc to make scoping cheap (that's literally why they exist)
- Give them a scoped target list (the standalone verifier and the crypto design are the cleanest bounded units)
Fit risk: OSTIF historically prioritizes critical / widely-deployed infrastructure (OpenSSL, Kubernetes, git). A small alpha tool may not be top-of-queue and may need to bring its own funder — which is exactly why the Sovereign-Tech-via-OSTIF ask matters. Timeline: rolling intake; once funded, weeks to a few months; total wall-clock often several months incl. fundraising. Confidence: high on the model; the open question is whether they take an alpha solo project without a pre-identified funder.
Verdict: Our existing posture is exactly what they reward. Lower friction than OTF/OSTIF and a good fit for a solo maintainer who'd benefit from mentorship plus money. Caveat: it's primarily a cohort-based maintainer security-uplift + education program (~3 weeks) with funding — the money can go toward an audit, but it complements rather than replaces an independent third-party one.
- Apply via the Secure Open Source Fund page:
github.com/open-source/github-secure-open-source-fund[RE-VERIFY URL + that a cohort is open] - Prepare the three deliverables: a written application about the OSS work, an interview, and a 45-second video pitch
[RE-VERIFY format] - Lead with the security-maturity signals — these are real and already in-repo:
make verify(ruff +mypy --strict+ pytest at ~85% coverage), CodeQL, pip-audit, Dependabot, SHA-pinned CI, signed-provenance + SBOM releases - Frame a clear security-improvement outcome: "fund and complete an external crypto/security review so the alpha caveat can be removed"
- Note: a next cohort was mentioned around September
[RE-VERIFY 2026 cohort dates]
Fit risk: Low on posture; the only caveat is that it's an uplift program, so pair it with OTF/OSTIF for the deep independent review. Timeline: cohort-based; weeks-to-months to cohort start, then ~3 weeks. Confidence: medium-high.
Verdict: EUR 5k–50k grants, and NGI Zero grantees can access free security AND accessibility audits via the coalition — which would bundle both reviews we need (roles 1 and 2). But the broad door is currently shut.
- Check
nlnet.nl/funding.htmlfor currently-open themes[RE-VERIFY open calls] - Note the flagship NGI Zero Commons Fund's 13th and FINAL call closed 2026-06-01 — that broad door is closed
- The only funds open mid-2026 — NGI Taler (privacy-preserving payments) and NGI Fediversity (federated hosting) — are poor thematic matches for a tenant evidence tool; the next deadline for those pilots is 2026-08-01
[RE-VERIFY] - Remember the audit support is reserved for projects already NGI-funded, so there's no audit without first winning a grant
- Action: monitor for a future broad/Commons-style NGI0 successor call rather than forcing a fit into Taler/Fediversity now (NLnet's application is famously lightweight, 1–2 pages, so it's cheap when a real fit opens)
Confidence: high on facts, medium-to-low on fit for the currently-open themes.
Verdict: Our license satisfies their rule (AGPL-3.0-or-later + Apache-2.0 verifier subset), but they fund critical base infrastructure with a EUR 50k+ minimum, and a single-maintainer alpha end-user app almost certainly does not clear that bar directly. The realistic value is as the money behind an OSTIF Bug Resilience engagement.
- Do not lead here as a direct applicant — route via OSTIF (see §2.2) and ask OSTIF to tap the Bug Resilience Program
- If you still want to check direct eligibility:
apply.sovereigntechfund.de(expect the EUR 50k+ / base-infrastructure bar)[RE-VERIFY] - Watch the Fellowship (2026 cohort closed 2026-04-06) and Standards (pilot closed 2026-05-19) for future rounds
[RE-VERIFY next windows]
Confidence: high on facts, low on direct fit.
Verdict: On paper an excellent audit-and-remediate model. In practice it has been on indefinite hiatus since the 2020 restructuring and is NOT accepting applications. Included only because it's commonly named.
- Do not apply — confirm status first if you must:
mozilla.org/en-US/moss/secure-open-source/[RE-VERIFY hiatus still in effect] - Monitor the Mozilla Technology Fund (thematic/cohort, historically not a general security-audit fund) in case SOS is revived
Confidence: high that it is on hiatus.
Use these for breadth and warm-up — especially the local web app + packet.html
attack surface — and reserve the deep crypto review for an OTF/OSTIF-facilitated
or paid specialist firm. Most pro-bono channels do org-level / vulnerability-assessment
work, not primitive-level review of Ed25519 / X25519 / ChaCha20-Poly1305 / RFC 3161.
- OSTIF as facilitator/matchmaker — covered in §2.2; it doubles as a pro-bono project-management channel even before we hold a grant.
- Consortium of Cybersecurity Clinics (university student clinics, Google-funded
network) —
cybersecurityclinics.org/about/our-members/. Contact a clinic directly (ideally nearby or one that does application-level work). Frame habitable as a public-interest tool for at-risk tenants. Expect a vuln/risk assessment from supervised students, not a deep crypto audit. Best as a cheaper complementary review or warm-up.[RE-VERIFY a member clinic is taking external orgs] - OWASP chapters + volunteer researcher communities —
owasp.org. Post baseline B1 and the standalone verifier as a concrete, bounded review target to a local chapter / relevant project list. Treat output as advisory, not an audit. - Academic applied-crypto / security labs (e.g. Citizen Lab and university groups) —
citizenlab.ca. No formal intake; opportunistic, topic-driven, long lead times, but potentially deep and citable at no cash cost. The chain-of-custody + RFC 3161 timestamping + verifier story is the most academically interesting angle — lead with that in cold outreach. Speculative; not a reliable channel.
Reusable narrative. Trim/expand per program; swap in OTF's at-risk framing, GitHub's security-uplift framing, etc. Replace every
[…]and re-verify program-specific facts before sending.
Project: habitable — github.com/ChelseaKR/habitable (public, AGPL-3.0-or-later; verification subset additionally Apache-2.0). Maintainer: Chelsea Kelly-Reif. Independent, unfunded, single-maintainer (bus factor 1). Currently v0.2.0, alpha.
The problem we're asking you to help validate. Tenants documenting habitability problems — mold, no heat, leaks, pests — produce evidence that a landlord has every incentive to dispute, and that a tenant may face retaliation for collecting. habitable is an offline-first, end-to-end-encrypted tool that turns those photos, videos, and notes into tamper-evident records: sealed originals with SHA-256 fixity, a hash-linked chain of custody, Ed25519 signatures, and RFC 3161 trusted timestamps (with archive re-timestamping), all checkable by a standalone verifier. The security is the product. We are not asking you to take that on faith — we are asking for an independent review to validate it.
Who it protects. People under threat of retaliation from a more powerful party
(a landlord), who often cannot afford a lawyer and whose evidence is only as good as
its integrity. The threat model treats the adversary (a landlord, or anyone who
can reach the device, the relay, or the timestamp authority) as the design center —
see docs/threat-model.md.
Why it's worth public-interest funding. Fully FOSS and reproducible; AGPL-3.0-or-later
with an Apache-2.0 verification subset specifically so courts and legal-aid orgs can
embed the verifier; synthetic-data-only evaluation (habitable demo) so no real
tenant data is ever needed to review it; documented privacy posture (DPIA-style
statement, docs/privacy.md). It serves an at-risk, under-resourced
population and asks nothing of them in return.
Scope of the audit requested. An independent security + cryptographic review that validates the stated claims and probes the attack surface:
- Cryptographic design & implementation — ChaCha20-Poly1305 vault-at-rest with a scrypt-wrapped key; Ed25519 signing; X25519 sealed-box E2E sync; SHA-256 fixity; correct, standards-conformant RFC 3161 timestamping and archive re-timestamping.
- Tamper-evidence guarantees — can a sealed original, a custody chain, or a timestamp token be altered such that the verifier still reports "intact" (e.g. undetected backdating)?
- The standalone verifier — its trust assumptions and failure modes on hostile input (a deliberately bounded, high-value target).
- Local web app +
packet.htmlattack surface (this part is also a good fit for a complementary/cheaper review). - Sync/relay confidentiality and the documented metadata exposure.
- Validation of the residual risks the maintainer has already enumerated in threat-model baseline B1 — confirm or refute them independently.
Budget sketch. [Tailor: for OTF the audit is contracted/paid by you; for OSTIF we're seeking facilitation + a funder; for GitHub SOS we'd apply cohort funding + program resources toward this review.] Indicative market cost of a focused crypto +
app audit of a project this size is roughly $30k–$200k (OSTIF's cited range);
our scope is deliberately bounded (one maintainer, a small, well-documented
codebase, a frozen threat model, and synthetic test data) to keep it toward the
lower end. [RE-VERIFY pricing before quoting a number.]
Project maturity / why this is a cheap, ready engagement (evidence, not claims). We treat our own assurance as committed artifacts, mirroring how the tool treats a photo's hash and timestamp:
make verifyreproduces the full merge gate: ruff,mypy --strict, and pytest with property-based convergence tests and tamper-detection tests against clean, altered, and chain-broken fixtures, at ~85% coverage (see theMakefile).- CodeQL, pip-audit, and Dependabot run in CI; CI actions are SHA-pinned;
releases ship signed provenance + an SBOM (
.github/workflows/). - A frozen, hash-pinned threat-model baseline B1
(
docs/audits/threat-model-baseline.md) so the reviewer and maintainer work from the same document and any later change is detectable. - A reviewer onboarding doc (
docs/audits/onboarding.md), a DPIA-style privacy statement (docs/privacy.md), an audit-as-artifact record (docs/audits/README.md) that already lists internally-found-and-fixed issues, and a private vulnerability reporting path (SECURITY.md). - An honest self-review stating plainly that no third-party audit has yet been
performed and that no novel cryptography was written (well-reviewed primitives
via the
cryptographylibrary).
What success looks like for us. The alpha caveat is gated on exactly this review (it's a stated precondition on the v1.0 gate). A completed, publishable audit — and the remediation it drives — is what lets us responsibly tell tenants and legal-aid orgs the tool can be relied on.
Contact. Chelsea Kelly-Reif — ckellyreif@gmail.com — github.com/ChelseaKR/habitable.
Keep this costed and ready so a funder/OSTIF "yes" converts to an award in days. Send to specialist firms (see §6).
- About the project. habitable, v0.2.0, alpha; FOSS (AGPL-3.0-or-later, Apache-2.0 verifier subset); single maintainer; github.com/ChelseaKR/habitable. Security is the product; users are at-risk tenants.
- Why now. Removing the alpha caveat is gated on this review. We have a frozen, hash-pinned baseline (B1) and reviewer onboarding ready, so scoping is cheap.
- In-scope (priority order).
(a) Cryptographic design + implementation: ChaCha20-Poly1305 at rest (scrypt-wrapped
key), Ed25519, X25519 sealed-box sync, SHA-256 fixity, RFC 3161 timestamping +
archive re-timestamping.
(b) Tamper-evidence: any path to alter a sealed original / custody chain / timestamp
while the verifier still says "intact" (undetected backdating especially).
(c) Standalone verifier (bounded, high-value).
(d) Local web app +
packet.htmlsurface. (e) Sync/relay confidentiality + documented metadata exposure. - Out of scope. Real tenant data (none exists; evaluate on
habitable demosynthetic data only). Roles 3–4 (legal framing, pilot) are handled elsewhere. - Reference materials provided. Threat model + baseline B1; onboarding doc; privacy statement; architecture + evidence-method docs; the full source.
- Deliverables. A written report with severity-rated findings and reproductions
on synthetic data; a remediation discussion; and a re-test of fixes. We intend
to publish the report in
docs/audits/(raise any constraints). - Engagement model. Fixed scope; coordinate via the private path in
SECURITY.mduntil fixes ship. - Timeline. Target an engagement window of
[N] weeks; we can start within[N] daysof award. - Commercials. Please quote fixed price + estimated effort; note any
FOSS / non-profit / public-interest rate. Funding is
[OTF-contracted / OSTIF-facilitated / grant-funded — specify].[RE-VERIFY firm pricing before committing a budget line.] - Selection criteria. Demonstrated applied-cryptography depth (primitive + protocol review), public open-source audit track record, and willingness to publish.
Validate crypto depth per engagement during scoping; confirm current intake and
pricing — none of the pricing below is verified. [RE-VERIFY all]
| Firm | Why it fits habitable | Note |
|---|---|---|
Trail of Bits — trailofbits.com/services/software-assurance/cryptography/ |
Dedicated cryptography service + open-source program; top-tier for our primitive/protocol set | Premium pricing |
Cure53 — cure53.de |
Deep, rigorous crypto + web-app audits; strong public-report reputation; fits both the crypto design and the packet.html/web surface |
Premium pricing |
Radically Open Security — radicallyopensecurity.com |
World's first non-profit security consultancy; crypto analysis + code audit; mission-aligned (channels ~90% of profit to NLnet); often more affordable | Strong budget shortlist pick |
Quarkslab — quarkslab.com |
Research-driven; already executes OSTIF-facilitated FOSS audits; crypto/RE depth | Plugged into the OSTIF pipeline |
NCC Group — nccgroup.com |
Dedicated cryptography-services practice; long public open-source audit history | Enterprise pricing/process |
Shielder / Ada Logics — shielder.com |
Smaller specialist FOSS-audit firms that take OSTIF engagements; better budget/scale match for a small project; still publishable | Confirm advanced-crypto depth in scoping |
Research dated 2026-06-17; the assistant's training cutoff predates it, so these were re-verified by live search then — but verify again at send time.
- OTF Security Lab — that
apply.opentech.fund/security-lab/project-side intake is open; current queue/turnaround; that we're not confusing it with the March 2026 auditor-recruiting RFP. - OSTIF —
ostif.org/get-an-audit/live;contactus@ostif.org(and the sponsorship/info@contact) current; whether they'll take an alpha solo project without a pre-identified funder. - GitHub Secure Open Source Fund — that a cohort is open; the application format (written + interview + 45s video); the 2026 cohort dates (~September?).
- NLnet — confirm the Commons Fund final (13th) call is closed (2026-06-01); that Taler / Fediversity are the only open funds (deadline 2026-08-01); watch for a successor broad NGI0 call.
- Sovereign Tech — EUR 50k+ minimum / base-infrastructure bar still in effect; next Fellowship / Standards windows; that Bug Resilience funds audits via OSTIF.
- Mozilla MOSS / SOS — confirm still on hiatus / not accepting applications.
- Paid firms — current intake + pricing for every shortlisted firm; any FOSS/non-profit rate.
- Pro-bono — that a Consortium of Cybersecurity Clinics member is taking external public-interest orgs; reachable OWASP chapter / researcher community.
Honesty note to carry into every pitch. habitable is alpha, unfunded, and bus-factor-1, and it protects domestic US tenants rather than the repressive-environment users some of these programs are framed around. Say so. The defensible, accurate frame — at-risk users facing retaliation, a privacy/tamper-evidence tool whose stated claims we want independently validated — is both true and our strongest argument.