Attestd: A Complete Project History and Postmortem

Project: Attestd Domain: attestd.io Founder: Robert Marshall Company: Marshall Digital Solutions Ltd Active period: Approximately February to October 9, 2026 Final status: Decommissioned Outcome: Technically operational product, commercially unsuccessful venture


Executive Summary

Attestd was an independently developed software-security intelligence platform built to serve a future in which autonomous systems and AI agents increasingly make software and infrastructure decisions without human intervention.

The project began as an attempt to make vulnerability information easier for software to consume. It evolved into a broader security intelligence API combining CVE intelligence, active exploitation status, software supply-chain compromise detection, package-name integrity, and AI-hallucinated package detection.

Its defining proposition was simple: a developer or autonomous system should be able to identify a piece of software, submit its name and version, and receive a deterministic, machine-readable assessment of known security risks without operating a vulnerability scanner, maintaining multiple threat-intelligence integrations, or relying on an LLM to interpret security advisories at runtime.

Attestd was built by a solo founder while employed full-time as a systems administrator in Canada, without venture capital, a dedicated engineering team, or a substantial marketing budget.

During its lifetime, the platform progressed from an experimental CVE lookup concept to production infrastructure with hundreds of supported infrastructure products, hundreds of thousands of monitored npm and PyPI packages, publicly distributed Python and JavaScript SDKs, MCP servers, agent framework integrations, automated security intelligence ingestion, and a public cryptographically signed detection ledger.

By the middle of 2026, Attestd was recording real software supply-chain incidents and demonstrating that its ingestion pipeline could surface known compromises ahead of mainstream security press coverage.

The engineering achievement, however, did not translate into meaningful market penetration.

On October 9, 2026, Attestd was decommissioned.

The founder’s final assessment was that the project failed primarily because it did not acquire sufficient market share before established cybersecurity companies expanded into the same problem space, progressively eliminating the differentiation that had justified Attestd’s existence.

The project demonstrated the feasibility of a useful security data abstraction. It did not establish an enduring commercial position from which to monetize that abstraction.

That distinction is central to understanding both Attestd’s history and its failure.

1. Origins: February to March 2026

The initial problem

Attestd emerged from a practical observation about the cybersecurity information ecosystem.

Software vulnerabilities are documented across numerous sources. The National Vulnerability Database, MITRE CVE records, vendor advisories, CISA’s Known Exploited Vulnerabilities catalog, and ecosystem-specific databases all contain useful information, but they serve different purposes and expose different structures.

They are primarily designed around the publication and consumption of security information.

A person can read a vulnerability advisory, interpret the relevant versions, consider the severity, identify whether exploitation is occurring, and decide what action to take.

A software system needs something considerably more precise.

It needs to know whether a particular version of a particular product is affected, what the operational risk is, how reliable that conclusion is, and whether it should escalate the result to another component.

Raw advisories do not always provide a straightforward answer.

Product names are inconsistent. A single vendor may have multiple CPE namespaces. Version constraints may overlap. Scores may be absent or misleading. Some records are poorly enriched. Exploitation status may come from a separate database altogether.

The initial Attestd concept was to resolve these complications into a structured response.

Rather than make every consuming application interpret the underlying security ecosystem, Attestd would do the interpretation once and provide the result through an API.

The premise was particularly compatible with the founder’s background in infrastructure operations.

Attestd was conceived less as another application for cybersecurity analysts and more as a dependable infrastructure service.

The first architectural direction

The early product concentrated on CVE intelligence.

A client would supply a software product and version. Attestd would normalize the identity, examine available vulnerability records, determine applicability, and return a structured conclusion.

The early engineering challenges were not glamorous, but they were consequential.

Correctness required handling inconsistent version strings, merging duplicate or overlapping vulnerability ranges, interpreting CPE mappings, and distinguishing an absence of known vulnerabilities from an absence of trustworthy information.

This led to an important principle:

Unsupported or insufficiently understood software must not be represented as safe.

The distinction between “no known risk” and “unknown risk” would remain an important part of Attestd’s design philosophy.

Another principle was determinism.

The same software version, evaluated against the same underlying intelligence, should produce the same security conclusion. An AI model might assist in preparing, classifying, or reconciling information during ingestion, but an ordinary API request should not depend on fresh generative reasoning.

This was both a technical decision and an early statement of product identity.

Attestd was being designed as infrastructure that could be relied upon, not as an AI chatbot offering security opinions.

The first major expansion

March brought a significant change in scope.

On March 24, the project recorded the compromise involving LiteLLM versions 1.82.7 and 1.82.8, in which malicious code was introduced through a software supply-chain attack.

The important feature of this incident was that traditional CVE intelligence alone did not adequately describe the problem.

A package could have no applicable CVE while still being actively malicious.

Consequently, Attestd’s design expanded beyond vulnerability assessment.

It began maintaining a separate software supply-chain integrity signal.

This distinction became foundational. A package might legitimately return no known CVE risk while independently being identified as compromised.

The two facts could coexist without contradiction.

Attestd would eventually make this distinction one of its principal selling points.

2. April 2026: From a CVE API to a Security Perception Layer

The April discussions marked a conceptual maturation of the project.

The question was no longer merely how to expose vulnerability information through a convenient API.

It became a question of how increasingly autonomous software should perceive security conditions in the environments it operates within.

Traditional security tools are often organized around discrete activities.

A developer initiates a scan. A CI/CD pipeline checks a dependency. A security professional reviews a dashboard. A compliance system generates a report.

Attestd’s proposed consumer was different.

An autonomous agent could decide to install a package, introduce a dependency, provision a service, upgrade software, or perform some other action without requiring a human to approve each step.

Such systems would benefit from security intelligence as an ordinary input to their decision-making.

This led to the project’s defining architectural distinction:

Attestd was a sensor, not a gate.

It would report the security state of software. It would not attempt to own the consuming application’s policy, authorization, or execution decisions.

The customer retained responsibility for deciding whether a particular condition warranted allowing, warning, reviewing, or blocking an action.

Attestd’s responsibility was to make the relevant security information available consistently, quickly, and in a form that software could interpret.

The machine-speed thesis

The broader argument developed during this period was that autonomous systems would fundamentally change the expected speed of security decision-making.

Human-oriented reports, dashboards, and advisory documents operate on human timescales.

Autonomous software can make and execute decisions much faster.

If autonomous systems were going to become meaningful participants in infrastructure operations, the security intelligence they consumed would need to operate within their own workflows.

The resulting thesis was that machine-speed consumers require machine-speed security information.

This was not an argument that AI models should independently invent risk assessments.

Quite the opposite.

The more consequential and autonomous these systems became, the greater the need for stable, structured, verifiable information that could be consumed without probabilistic reinterpretation during every request.

Attestd would increasingly be framed as the security perception layer for autonomous infrastructure.

That terminology was useful for explaining the long-term vision, although it also created a marketing difficulty: potential buyers did not necessarily recognize “security perception layer” as an established category.

The problem of creating a new category while simultaneously introducing a new product would follow Attestd throughout its commercial life.

3. May 2026: Building a Real Production Service

May was a substantial engineering and operational milestone.

By May 27, internal records described a production system with more than 100 supported CVE products, initial npm and PyPI supply-chain monitoring, published SDKs, framework integrations, and automated package qualification.

Attestd was no longer an architectural concept.

It was a functioning service.

Infrastructure and reliability

The production architecture evolved toward a separation of concerns intended to protect reliability without requiring enterprise-scale expenditure.

Neon PostgreSQL served as the independent database layer. Railway became the primary API compute platform, with Render providing secondary compute capacity. Cloudflare provided the external routing and failover layer.

This architecture reflected the founder’s operational background and concern about single-provider failure.

A provider outage in May reinforced the decision to separate database infrastructure from compute infrastructure and establish redundancy.

By the June launch preparation period, the project reported 99.995% uptime over a 90-day window. Its internal reliability notes described the redundant architecture as having been exercised by a real provider failure.

For a small, independently financed software product, this was a significant commitment to operational quality.

It also illustrated a recurring tension in Attestd’s development.

The engineering standard was being established as though the service were already a dependency of serious production customers, even though widespread customer adoption had not yet been achieved.

SDKs and developer integrations

The Python SDK, published on PyPI, and JavaScript SDK, published on npm, provided direct programmatic access to the service.

The JavaScript SDK was deliberately lightweight, supporting both ESM and CommonJS, with no external dependencies and package provenance safeguards.

The product also expanded into tools used by AI developers.

Integration work included LangChain, CrewAI, AutoGen, and the Model Context Protocol.

The MCP strategy became particularly important.

Attestd offered both a locally launched stdio MCP server and a remotely accessible HTTP MCP server.

The intention was to make security intelligence available to agent development environments without requiring developers to write a new integration from scratch.

This was an early expression of what became the project’s principal distribution theory.

If autonomous systems were the intended consumers, Attestd needed to become easy to discover and call from the tools those systems already used.

Automated package monitoring

The supply-chain ingestion architecture also became increasingly automated.

It drew intelligence from public sources, particularly OSV and malicious-package advisory records, while incorporating package download information and AI-assisted name-integrity checks into qualification workflows.

A six-hour ingestion cycle became part of the operating system.

The project recorded new intelligence, classified relevant software, and reported developments through automated operational channels.

This was important because Attestd could not realistically scale security monitoring through manual package curation alone.

Automation was not merely intended to reduce operational expenses.

It was required for the product’s underlying promise to make sense.

4. June 2026: Technical Scale, Evidence, and the Public Thesis

June was arguably Attestd’s strongest period of technical and conceptual progress.

The system expanded dramatically in coverage, developed a more credible public evidence trail, refined pricing, and prepared for a major developer-community launch.

The coverage expansion

The difference between late May and the June launch preparations was substantial.

MetricMay 27 recordJune launch-preparation record
CVE-covered products100+350+
PyPI supply-chain packagesApproximately 5126,405
npm supply-chain packagesApproximately 70230,463
Combined monitored packagesApproximately 121More than 256,000

The expansion did not mean every package had received individual human analysis.

It reflected the development and scaling of automated intelligence ingestion and qualification systems.

Nevertheless, it changed Attestd’s practical usefulness.

An API covering a few dozen packages was an interesting demonstration. A service monitoring hundreds of thousands of packages could plausibly be integrated into real developer workflows.

The project also maintained a more selective CVE product qualification process, intended to avoid inflating product counts through poor-quality mappings or unusable vulnerability records.

Coverage was increasingly treated as a combination of breadth and reliability rather than a simple marketing number.

The detection record

Attestd’s operating records documented a sequence of software supply-chain incidents identified by its ingestion pipeline.

These included:

DateRecorded event
March 24LiteLLM compromised releases
April 27elementary-data malicious-package incident
May 12TanStack and Mistral package compromises
May 14node-ipc incident
May 19Compromises involving the @antv ecosystem
May 22Laravel Lang git tag rewriting attack
June 1codexui-android targeting developer credentials
June 12–16Malicious PostCSS-related packages
June 16–17Mastra ecosystem compromise and typosquat activity
June 24–28VS Code Fake Font campaign and related malicious packages

The contemporaneous records claimed that several events were represented in Attestd’s intelligence before major security publications reported them.

Examples included the TanStack and Mistral detections, recorded roughly two and a half hours before BleepingComputer’s coverage, and the VS Code Fake Font packages, recorded several days before later reporting.

These were meaningful observations, but an important distinction applies.

Detecting a compromise before a news article is published does not necessarily mean independently discovering an attack before any upstream intelligence source.

Attestd primarily synthesized, qualified, and distributed security intelligence. Its strength was the automated collection and delivery of actionable known conditions.

That was a legitimate capability without needing to characterize every ingestion event as original threat discovery.

The public detection ledger

A major strategic discussion in June concerned credibility.

A newly established security API operated by a solo founder faced an obvious problem: why should anyone trust its security conclusions?

This challenge extended beyond technical correctness.

Even if Attestd had detected real incidents, prospective customers had little reason to accept unsupported claims from an unfamiliar company.

The solution was a public detection ledger.

Attestd established a GitHub repository recording security events, detection times, source publication times, and other evidence. Entries used cryptographically signed commits to create an externally inspectable operating history.

The ledger was also exposed through the website.

This was an attempt to convert Attestd’s claims into inspectable evidence.

It reflected one of the strongest decisions made during the project’s lifetime.

Rather than simply insist that the product was trustworthy, the company would provide a mechanism through which others could investigate that trustworthiness.

The ledger also demonstrated an understanding of an important reality in cybersecurity: credibility is part of the product.

The Autonomous Infrastructure Thesis

On May 28, 2026, Attestd had published The Autonomous Infrastructure Thesis.

By June, this document had become the project’s principal explanation of why it ought to exist.

It drew a distinction between traditional software automation and the emerging behavior of AI-driven systems.

The thesis discussed how agentic systems re-query software information, make decisions outside fixed deployment schedules, and require consistency from their external information sources.

It argued that the proliferation of capable autonomous systems would create demand for standardized security intelligence that could be consumed directly.

The purpose of Attestd was to provide that intelligence without adding nondeterministic generative reasoning to the request path.

In hindsight, this may have been one of the project’s clearest technical insights.

It was also an insight that larger cybersecurity companies would be capable of recognizing independently.

Being correct about a market’s direction did not guarantee that Attestd would capture the resulting demand.

5. June to July: Pricing, Positioning, and the Attempt to Find Distribution

As Attestd matured technically, attention increasingly turned to how it could become a sustainable business.

The early pricing structure had been based primarily on API volume.

Historical records describe an initial Free tier with approximately 1,000 monthly calls, a $19 Starter tier with 25,000 calls, and a $79 Pro tier with 200,000 calls.

This model was straightforward to meter, but it created a commercial problem.

Attestd’s underlying data sources were largely public. Prospective developers could reasonably compare a paid API call against directly consuming free databases such as OSV.

The project needed customers to understand why the normalized intelligence was valuable, not merely that the endpoint was convenient.

The pricing rethink

In June, the pricing strategy was reconsidered.

The central question was whether Attestd should charge primarily for queries or for the ongoing confidence and assurance that its intelligence could provide.

One proposed direction was to make ordinary API consumption inexpensive or effectively unlimited under fair-use constraints, while reserving monitoring, webhooks, change feeds, environment-specific keys, and service commitments for higher tiers.

The rationale was that customers should pay for additional value rather than simply to remove artificial usage restrictions.

Embedding introduced a separate issue.

A company using Attestd internally was commercially different from a company integrating Attestd into a product used by thousands of its own customers.

The original proposed answer was a separate negotiated Platform tier for embedded and reseller usage.

The pricing strategy continued to evolve over subsequent months.

By the later product period, the public offer was organized around approximately:

TierIntended offer
Free5,000 calls per month
Solo$19.99 per month, up to 250,000 calls
Team$99.99 per month, up to 2 million calls
PlatformCustom commercial terms

Team-level capabilities included scoped keys, webhook support, and scheduled intelligence digests, although not every proposed feature was generally available.

In the later outreach discussions, the Team tier’s embedding rights were also reconsidered and expanded to reduce friction for prospects considering Attestd as a component of their own products.

These revisions reflected a gradually improving understanding of the intended buyer.

Attestd was not primarily selling vulnerability lookups to individual enthusiasts.

Its most meaningful opportunity was becoming an embedded infrastructure dependency inside another company’s software.

The competitive positioning problem

Attestd’s intended category was difficult to communicate.

It was not a conventional vulnerability scanner.

It did not scan customer repositories or operate as a security dashboard.

It did not aim to replace Snyk, Sonatype, Socket, JFrog, or other established cybersecurity platforms.

Its preferred positioning was upstream of those systems: downstream of public intelligence sources, but upstream of the applications requiring security conclusions.

This was architecturally coherent.

Commercially, however, it was challenging.

Customers understood security scanners, dependency management, package protection, and repository security because those were established purchasing categories.

They did not necessarily understand why they needed a separate “security perception layer.”

Explaining the product often required explaining the future operating model of autonomous systems first.

That increased the cost of every sales conversation.

The project attempted to simplify the message to a single practical promise: one request returns the relevant security facts for software that a system is about to select, install, deploy, or execute.

Even then, the prospect could ask why the same function could not be supplied by an existing security vendor.

That question would become increasingly difficult to answer as the market evolved.

6. July 2026: The Planned Launch and the Search for a Beachhead

A significant portion of the June development and marketing work was organized around a proposed Hacker News launch.

The launch preparation became unusually thorough.

It included a public About page, legal documentation, technical guides, SDKs, MCP directory submissions, published case studies, a verifiable detection ledger, developer-focused landing pages, and prepared responses to likely criticism.

The date moved several times, eventually reaching a July 8 target.

The surviving master preparation task was later marked cancelled, so the archive does not establish that the planned launch occurred as originally intended or delivered meaningful traction.

What it does establish is the amount of preparation devoted to an opportunity for developer discovery.

The distribution strategy also included relevant developer communities, technical articles, GitHub repositories, package registry listings, and search traffic.

The project explicitly preferred technical credibility over aggressive marketing.

This was consistent with the founder’s available resources and professional temperament.

Attestd had no large sales organization, no significant advertising budget, and no established cybersecurity brand.

It needed developers to discover the product, recognize its usefulness, and begin integrating it.

The autonomy argument becomes more concrete

The July strategic material connected Attestd’s thesis to reports of increasingly autonomous offensive cyber activity.

A documented AI-driven ransomware operation involving an older, already known vulnerability became an example used in Attestd’s positioning.

The argument was not that Attestd could prevent every such attack.

It was that autonomous defensive and operational systems needed timely structured intelligence about known vulnerabilities and exploitation conditions.

That was a defensible product narrative.

It also reinforced that Attestd’s opportunity depended on the adoption trajectory of autonomous software.

The platform was being built ahead of a category whose ultimate demand and purchasing practices were still developing.

7. July to August: Technical Success Without a Corresponding Market Position

By late July, the recorded coverage had reached approximately 356 CVE-covered products, 27,459 PyPI packages, and 237,601 npm packages.

The service had become technically substantial.

It included independently hosted API infrastructure, intelligence ingestion, SDK distribution, MCP accessibility, public documentation, technical content, and operational monitoring.

It was also designed to remain inexpensive enough for the founder to sustain without institutional investment.

Monthly operating costs were generally discussed in the approximate range of CAD $150 to $250 during the broader project period, though the archive does not establish a complete final accounting.

Attestd’s original ambition was considerably larger than its operating budget.

At one stage, the founder envisioned developing the company into an asset capable of achieving an acquisition valuation of approximately $20 million to $25 million within two to three years.

This was a strategic ambition, not a demonstrated valuation.

The plan was to remain lean, avoid large headcount, and eventually prioritize sales capability over further engineering expansion.

The underlying bet was that security infrastructure embedded in multiple growing software platforms could achieve valuable recurring revenue and become attractive to an acquirer.

The difficulty was getting from a functioning API to a meaningful number of embedded users.

The product increasingly had the infrastructure necessary to support adoption, but infrastructure alone did not create it.

In parallel, established cybersecurity vendors already possessed customer relationships, distribution channels, security reputations, and integration ecosystems that Attestd had not yet built.

Even where Attestd offered a cleaner or more directly consumable abstraction, larger vendors had strong incentives to expand into the same autonomous-agent security use cases.

Attestd was entering a market where the underlying opportunity was becoming visible to companies with considerably more resources.

8. September 2026: The Final Commercial Strategy

September brought perhaps the clearest articulation of Attestd’s realistic commercial path.

The emphasis shifted toward finding a small number of genuine production integrations.

Instead of broadly promoting a security API and hoping developers would adopt it, the project would target software companies for whom Attestd could become an embedded component.

The proposed Platform Starter / Founding Embed program sought to win three to five authentic production integrations.

This was a meaningful refinement.

It recognized that an integration announcement, a demo, or an API key registration did not constitute product-market fit.

A genuine integration needed to operate inside a real customer workflow, influence an actual software decision, remain in use after onboarding, and demonstrate economically sustainable usage.

The intended integration point was explicit.

An autonomous system would identify software it intended to select, install, upgrade, load, deploy, or execute. It would call Attestd for available vulnerability, supply-chain, and name-integrity intelligence. The consuming system would retain control over enforcement.

The same pattern could potentially serve coding agents, package managers, registries, sandboxes, remediation tools, MCP infrastructure, and other automated systems.

The proposed commercial experiment included a low-friction startup embedding offer. One internal candidate was approximately US$99 monthly for 100,000 assessments, although that figure was exploratory rather than proof of a universally launched pricing product.

This approach recognized a basic commercial truth.

A small company willing to put Attestd into production was more valuable than a famous prospect that expressed interest without integrating.

Prospect research and outbound

The project developed prospect databases and examined potential customers or integration partners.

The discussion increasingly centered on platform owners and products that would encounter software trust decisions as part of their normal operations.

The value proposition was that these companies could consume Attestd’s maintained intelligence instead of building and operating a comparable data normalization and threat-monitoring system themselves.

The earlier September outreach discussions also prompted modifications to embedding rights.

The aim was to remove commercial restrictions that might prevent customers from incorporating Attestd into their own products.

However, there was no established evidence in the surviving record that the planned production embedding cohort had been successfully acquired and retained.

The commercialization strategy had become more precise, but time was becoming a more serious constraint.

9. September: The Agentic Software Factory

Attestd also became a significant experiment in the use of autonomous AI-assisted software development and operations.

The founder was simultaneously maintaining full-time employment and developing multiple products.

As a result, operational leverage became increasingly important.

The project developed a structured operating system using Notion, GitHub, Cursor, and Grok-based agents.

The roles eventually included research, project management, editing, maintenance, and a quality-assurance function.

Research agents were responsible for monitoring threat intelligence, ecosystem developments, competitive changes, and coverage gaps.

Maintenance agents examined code, dependencies, documentation, CI failures, and implementation tasks.

An editorial workflow handled technical content.

A project-management role coordinated work, reviewed bounded changes, and managed accepted pull requests.

The operational structure evolved throughout September.

By September 19, the separate Assurance agent had been retired, with its responsibilities absorbed into project management.

By September 23, a standing Cursor Project had become the standard implementation route for product coding.

By September 25, the project manager’s standing review and merge responsibilities had been formalized.

The operating system had become sophisticated enough to route work across several repositories, manage recurring investigations, perform code maintenance, and handle selected publication activities.

Security-sensitive change control

One feature of this system was particularly characteristic of Attestd.

Changes that might alter customer-facing security conclusions remained founder-gated.

Autonomous agents could investigate freely and implement bounded low-risk changes, but they could not independently rewrite detection semantics, vulnerability associations, compromise classifications, exploitation states, or other consequential risk judgments.

This distinction was deliberate.

A security product that allowed an autonomous maintenance process to silently change what it told customers about risk would undermine its own trust model.

Attestd therefore attempted to use AI automation without surrendering responsibility for the accuracy of security decisions.

This was a relatively advanced approach to operating a small software company.

It also exposed a painful commercial contrast.

The founder had developed increasingly sophisticated mechanisms for keeping the product healthy and extending its capabilities.

What had not developed at a comparable rate was an engine for acquiring customers.

The software factory could accelerate engineering, research, and maintenance.

It could not manufacture market demand.

10. The Competitive Landscape and the Erosion of Differentiation

From the beginning, Attestd faced competition from companies with existing positions in vulnerability management, dependency intelligence, package scanning, and software supply-chain protection.

Its early answer was to distinguish itself architecturally.

Sonatype and Snyk were generally associated with developer and security workflows. JFrog offered repository-level and artifact-management controls. Socket specialized in package and dependency risk, including behavioral analysis.

Attestd argued that its particular strength was a lightweight, read-only, deterministic intelligence API that could be inserted directly into an autonomous decision boundary.

That was a meaningful distinction at the time it was formulated.

But it was not necessarily a durable competitive barrier.

Established vendors could build agent integrations, publish MCP servers, expose simplified security APIs, expand package metadata, or incorporate machine-readable risk assessments into their existing offerings.

They also had advantages that Attestd could not reproduce quickly: recognized brands, established customer contracts, proprietary or enriched datasets, security research teams, existing purchasing relationships, and enterprise distribution.

The decisive problem was not that Attestd became incapable of performing its intended function.

It was that performing that function stopped being sufficiently distinctive.

A prospect who could obtain broadly comparable functionality through an existing vendor had less incentive to introduce a new security dependency from an unfamiliar solo-operated company.

Even when Attestd remained technically competitive in a narrow area, the cost of convincing a customer to trust and integrate it could be substantial.

The founder’s retrospective judgment is that established cybersecurity companies expanded into Attestd’s reason to exist before Attestd secured enough market share to defend that position.

The surviving project documents do not establish one particular competitor announcement as the single event that ended the company.

Rather, the failure reflects a broader compression of the commercial opportunity.

The original differentiation became less exclusive while the distribution problem remained unresolved.

11. October 2026: Decommissioning

Attestd remained operationally active immediately before its closure.

An October 8 competitor and ecosystem research record documents continued monitoring of registry developments, supply-chain incidents, agent sandboxing, and security-vendor capabilities.

That report found no individually decision-changing development during its immediate review window.

The project was still maintaining the operating processes built during September.

On Friday, October 9, 2026, the founder confirmed that Attestd had been decommissioned.

The decision closed a venture that had consumed approximately eight months of development, research, infrastructure work, architectural reasoning, commercial planning, and operational effort.

The final outcome was not the result originally envisioned.

Attestd did not become a widely embedded security intelligence provider.

It did not establish the market position necessary to achieve the desired commercial scale or justify the long-term acquisition ambition.

The available historical record does not support assigning a precise lifetime revenue figure, paid customer total, or complete financial loss. Those details should not be invented for the sake of a neat retrospective.

What is established is the commercial conclusion.

The project had failed to gain sufficient market share, and its remaining differentiating value was being absorbed by larger companies.

The founder chose to discontinue it.

12. What Attestd Actually Accomplished

A commercial failure does not require pretending the underlying engineering never happened.

Attestd produced a working system with several substantial achievements.

It demonstrated that public vulnerability and software supply-chain intelligence could be synthesized into a stable API contract appropriate for automated consumers.

It established a useful separation between traditional CVE risk and malicious-package compromise.

It made package-name integrity and AI-hallucinated dependency detection part of the same broader software trust problem.

It developed automated ingestion and qualification processes capable of maintaining coverage across large software ecosystems.

It delivered SDKs, MCP servers, agent integrations, documentation, and a resilient hosting architecture.

It accumulated an operating record of real supply-chain incidents and established a public detection ledger intended to make its historical claims independently inspectable.

It also served as a practical test of agent-assisted software engineering.

By the final months, a single founder could coordinate research, maintenance, documentation, development, and parts of deployment through structured autonomous workflows.

That organizational capability did not save the business.

But it was a meaningful technical and operational result.

The central achievement was that Attestd became a real product.

The central failure was that it did not become a necessary product for enough customers.

Those two statements can both be true.

13. Why Attestd Failed

The primary reason, as identified by its founder, was failure to acquire market share before better-established cybersecurity firms occupied the same commercial opportunity.

Several contributing mechanisms help explain that outcome.

Technical progress outpaced commercial validation

Attestd advanced rapidly in product scope, reliability, integrations, and supported software coverage.

The commercial customer base did not advance sufficiently alongside it.

This created an imbalance.

Each technical improvement strengthened the product, but did not necessarily strengthen its position in the market.

The engineering roadmap was substantially more mature than the evidence of recurring willingness to pay.

Distribution was treated as something that could compound organically

The strategy relied heavily on developer discovery, SEO, technical content, package registries, open integration surfaces, community launches, and eventually agent recommendations.

This made sense within the financial constraints of a bootstrapped solo venture.

It also meant that distribution could develop slowly even while engineering developed quickly.

An API may be technically excellent and well documented while remaining largely invisible to the people who could benefit from it.

Attestd never established the necessary distribution velocity before the competitive landscape changed.

The product entered an emerging category

The long-term target was autonomous infrastructure and agentic software.

That market was still developing, including its purchasing behaviors, integration standards, and expectations of built-in security.

Attestd could articulate why autonomous systems would need security intelligence.

But recognizing future demand is not the same as reaching an immediate buyer with an urgent budget and a willingness to integrate.

The project spent substantial effort preparing for the market it expected, without securing enough of the market that existed.

Trust was expensive to establish

Cybersecurity is a difficult market for an unknown supplier.

A new API that reports whether software is dangerous creates a consequential dependency.

Potential customers need confidence in accuracy, evidence provenance, reliability, maintenance, failure behavior, and long-term continuity.

Attestd addressed these problems seriously through documentation, redundancy, public identity, and a signed detection ledger.

But those mechanisms could not instantly produce the institutional credibility already held by established vendors.

The solo-founder model helped control costs and engineering velocity, but it also made the trust problem harder.

The differentiation was vulnerable to replication

Attestd’s strongest insight was architectural: security intelligence should be a lightweight, deterministic primitive consumed by machines.

However, the product did not possess an exclusive source of the underlying security information, an established distribution monopoly, or a large installed base.

Its synthesis and delivery mechanisms were valuable, but larger competitors could plausibly recreate many of the same capabilities.

The defensible commercial asset would likely have needed to become the installed base, workflow integration, brand trust, or an increasingly valuable proprietary intelligence capability.

Attestd did not accumulate those advantages quickly enough.

The founder had finite time

Attestd was developed alongside demanding full-time employment and other software ventures.

Its operating model used AI-assisted engineering to extract considerable output from limited founder time.

Nevertheless, software development, product management, security assurance, content, prospecting, partnerships, and sales all competed for the same attention.

The commercial strategy eventually became focused on genuine production embeddings, but this clarity arrived late in the venture’s short lifespan.

This is a contributing constraint, not a substitute for the founder’s stated reason for failure.

Even a larger time commitment would not automatically have produced market share.

14. The Most Important Lessons

A correct thesis is not a captured market

Attestd identified a real direction in the evolution of software security.

Autonomous software would need security information in forms that autonomous systems could consume.

That insight did not guarantee that a new company would own the resulting market.

Recognizing a transition is different from possessing an advantage in the transition.

An incumbent with existing customers may benefit more from a change than the startup that articulated it earlier.

Building a category creates an additional burden

Attestd attempted to introduce both a new service and a new way of thinking about how security intelligence should be consumed.

That made the product intellectually compelling but more difficult to sell.

A customer first had to accept the category, then understand the specific product, then trust the operator, then justify integrating another dependency.

Every additional step introduced friction.

Engineering velocity does not equal commercial velocity

Attestd’s development demonstrated considerable technical productivity.

AI-assisted development made it possible to build, test, document, and operate systems at a pace that would historically have required considerably more manpower.

But the ability to ship software did not make customers appear.

The constraints shifted from implementation capacity toward attention, credibility, differentiation, and distribution.

The venture became an example of how the barrier to creating competent software can fall much faster than the barrier to creating a successful software business.

A startup must turn temporary differentiation into durable advantage

Being earlier to a product capability can create an opportunity.

It does not automatically create a moat.

Attestd needed to convert its initial architectural advantage into adoption, partnerships, integration depth, recurring revenue, and eventually switching costs.

The final commercial strategy recognized that necessity.

The company did not complete the transition before competitors reduced the uniqueness of its offer.

A small operational footprint is not sufficient

Attestd was inexpensive to operate relative to its technical scope.

That was a valuable characteristic.

But low operating costs can extend the life of a business without changing whether that business has a defensible reason to exist.

The decision to decommission acknowledged that continued technical operation was not necessarily the best use of the founder’s resources.

15. What Was Left Unfinished

The most consequential unfinished task was commercial rather than technical.

Attestd had developed a plausible method of becoming an infrastructure dependency for other software companies.

It had not proven that method through a sufficient base of deployed, retained, paying integrations.

The Platform Starter program was an attempt to cross that threshold.

The product’s roadmap also included broader assurance features, event-driven monitoring, more complete coverage, continued SDK and MCP improvements, and further automation of threat research and quality control.

Some elements were implemented, some remained partial, and others existed as plans.

Their completion would have increased the scope or convenience of the platform.

There is no basis for assuming they would have reversed the commercial result.

At shutdown, the central unanswered question was no longer whether Attestd could be made technically capable.

It was whether enough customers would choose it over expanding incumbent alternatives.

The founder’s answer was that the remaining opportunity no longer justified continuing the venture.

16. Final Assessment

Attestd was an ambitious attempt to build a small but strategically positioned piece of infrastructure for autonomous software.

It began with an identifiable technical problem, developed a coherent architecture, and evolved into a functioning security intelligence service.

Its strongest aspects were the precision of its underlying data model, its separation of independent security signals, its automation-oriented architecture, its public evidence strategy, and its disciplined approach to security-sensitive change control.

Its weakest aspect was commercial adoption.

The venture did not establish a customer base or market position that could sustain its differentiation as larger competitors expanded into the space.

This was not a case of an idea never being implemented.

It was a case of a product being implemented without successfully capturing the opportunity on which its business depended.

The most important conclusion is that Attestd succeeded in building much of what it set out to build, but failed to become the company it needed to become.

The founder correctly recognized an emerging architectural requirement. The project demonstrated that requirement through working software. It then lost the commercial race to establish itself as a preferred provider before established companies made similar capabilities increasingly accessible.

On October 9, 2026, that experiment ended.

Epilogue

Attestd began with a premise: that the systems making decisions about software would increasingly need to understand security without waiting for a person to interpret it for them.

Over the following months, that premise became an architecture, then an API, then a production service, then an increasingly automated small software operation.

The project had real infrastructure, real code, real operational complexity, and a documented history of processing genuine security incidents.

But it did not secure the customers necessary to make those achievements commercially durable.

The market did not remain unoccupied while the product matured.

The opportunity was recognized by larger companies whose existing relationships and resources allowed them to approach the same problem from positions of strength.

Attestd reached the end of its commercial justification before it reached the scale originally envisioned.

Its closure was therefore a failure of the business thesis as executed, not a claim that the engineering work had been imaginary or worthless.

For posterity, the distinction matters.

The venture demonstrated how much a determined individual could build using modern software infrastructure and increasingly capable AI development tools.

It also demonstrated what those tools could not provide automatically: customers, distribution, trust, and a durable market position.

Attestd, 2026. Built, deployed, operated, and ultimately decommissioned.

The product worked. The business did not.


Archival References

The following surviving records provide useful primary documentation of the project during its operating life. They capture the project’s thinking and claimed results at the time they were written, rather than independently verified post-closure findings.

  • Chat Migration Dossier, May 27, 2026: Early production architecture, coverage, integrations, operating principles, and development roadmap.
  • Fable Strategic Analysis, June 2026: Trust, distribution, pricing, evidence, and competitive positioning.
  • Show HN Master Preparation Document: Technical capabilities, historical detection claims, infrastructure reliability, launch planning, and the autonomous infrastructure thesis.
  • Pricing Overhaul, June 2026: Evolution of the pricing model and early distinction between direct usage and commercial embedding.
  • Platform Starter and Production Embedding Program, September 2026: The final major commercial strategy aimed at genuine production integrations.
  • Attestd Hub: The autonomous operating structure, founder approval rules, development workflow, and operational responsibilities.
  • Ecosystem and Competitor Scan, October 8, 2026: One of the final surviving operational records before decommissioning.

Historical record prepared October 9, 2026.

docs / detection-ledger

Detection Ledger

A public commit record of supply chain compromise incidents registered by Attestd. Each incident lists packages, versions, how Attestd registered the signal, and OSV or advisory citations when available.

Entries are pushed to a public GitHub repository. Commit timestamps show when the ledger entry was published. This is an archive of what Attestd registered, not a claim that Attestd beat OSV or press.

Three entries are worth reading in full: VS Code Fake Font, flagged five days before JFrog's disclosure, the Shai-Hulud attack on TanStack and Mistral, flagged roughly two and a half hours before BleepingComputer's coverage, and the AsyncAPI re-compromise, caught by live monitoring rather than a lookup against the original advisory.

detection record

37763 incidents across npm and pypi. 36769 with an OSV citation (53 demo or sandbox entries hidden). 22526 verifiable within 24h of attestd_detected_at; 15237 commit lag; 29189 indexed from historical OSV.

Show all entries including demos

Commit lag and historical OSV index are independent counts. A 24h-verifiable row can still be indexed from historical OSV.

Full commit history and JSON artifacts: github.com/attestd-io/detection-ledger

verifiable within 24h

ledger_commit_at is within 24 hours of attestd_detected_at. That 24h window is a publication lag bound. It does not mean GitHub commit time equals detection time. Detection time is first_seen / attestd_detected_at. The GitHub commit publishes an already-recorded detection.

Showing 100 of 22526. Show all in this section · Full ledger.json

IncidentSourceRegisteredOSVCitation
ig-gox-2026-10-09
ig-gox
ig-gox
osv2026-10-09 10:01 UTCMAL-2026-17712link
wix-reg-poc-bypass-2026-10-09
wix-reg-poc-bypass
wix-reg-poc-bypass
osv2026-10-09 00:59 UTCMAL-2026-17711link
revine-2026-10-08
revine
revine
osv2026-10-08 22:43 UTCMAL-2026-17709link
pxnpm-2026-10-08
pxnpm
pxnpm
osv2026-10-08 22:43 UTCMAL-2026-17710link
wxwxtest-testrrrdd-2026-10-08
@wxwxtest/testrrrdd
@wxwxtest/testrrrdd
osv2026-10-08 21:58 UTCMAL-2026-17704link
css-overscroll-contain-2026-10-08
css-overscroll-contain
css-overscroll-contain
osv2026-10-08 21:58 UTCMAL-2026-17705link
testrrrd-2026-10-08
testrrrd
testrrrd
osv2026-10-08 21:58 UTCMAL-2026-17706link
wie888r-2026-10-08
wie888r
wie888r
osv2026-10-08 21:58 UTCMAL-2026-17707link
xblaxw-2026-10-08
xblaxw
xblaxw
osv2026-10-08 21:58 UTCMAL-2026-17708link
kafka-roller-2026-10-08
kafka-roller
kafka-roller
osv2026-10-08 21:28 UTCMAL-2026-17703link
kafka-helmsman-2026-10-08
kafka-helmsman
kafka-helmsman
osv2026-10-08 21:28 UTCMAL-2026-17702link
kxafunc-xbails-2026-10-08
@kxafunc/xbails
@kxafunc/xbails
osv2026-10-08 18:27 UTCMAL-2026-17701link
dransay-2026-10-08
dransay
dransay
osv2026-10-08 17:56 UTCMAL-2026-17700link
galicia-toolkit-error-master-2026-10-08
@galicia-toolkit/error-master
@galicia-toolkit/error-master
osv2026-10-08 17:26 UTCMAL-2026-17689link
galicia-toolkit-spa-build-config-2026-10-08
@galicia-toolkit/spa-build-config
@galicia-toolkit/spa-build-config
osv2026-10-08 17:26 UTCMAL-2026-17690link
galicia-toolkit-tag-manager-2026-10-08
@galicia-toolkit/tag-manager
@galicia-toolkit/tag-manager
osv2026-10-08 17:26 UTCMAL-2026-17691link
dransay-secrets-2026-10-08
@dransay/secrets
@dransay/secrets
osv2026-10-08 17:26 UTCMAL-2026-17699link
dransay-db-2026-10-08
@dransay/db
@dransay/db
osv2026-10-08 17:26 UTCMAL-2026-17695link
dransay-feature-flags-2026-10-08
@dransay/feature-flags
@dransay/feature-flags
osv2026-10-08 17:26 UTCMAL-2026-17696link
dransay-logger-2026-10-08
@dransay/logger
@dransay/logger
osv2026-10-08 17:26 UTCMAL-2026-17697link
bitwardne-cli-2026-10-08
@bitwardne/cli
@bitwardne/cli
osv2026-10-08 17:26 UTCMAL-2026-17693link
bitwardne-jslib-2026-10-08
@bitwardne/jslib
@bitwardne/jslib
osv2026-10-08 17:26 UTCMAL-2026-17694link
clickfix-npm-delivery-2026-10-08
clickfix_npm_delivery
clickfix_npm_delivery
osv2026-10-08 16:41 UTCMAL-2026-17692link
galicia-toolkit-core-2026-10-08
@galicia-toolkit/core
@galicia-toolkit/core
osv2026-10-08 16:40 UTCMAL-2026-17688link
tailwindcss-effects-kit-2026-10-08
tailwindcss-effects-kit
tailwindcss-effects-kit
osv2026-10-08 15:39 UTCMAL-2026-17686link
tailwindcss-ui-extensions-2026-10-08
tailwindcss-ui-extensions
tailwindcss-ui-extensions
osv2026-10-08 15:39 UTCMAL-2026-17687link
node-fetch-lite-2026-10-08
node-fetch-lite
node-fetch-lite
osv2026-10-08 15:39 UTCMAL-2026-17683link
node-gyp-runtime-2026-10-08
node-gyp-runtime
node-gyp-runtime
osv2026-10-08 15:39 UTCMAL-2026-17684link
system-driver-2026-10-08
system-driver
system-driver
osv2026-10-08 15:39 UTCMAL-2026-17685link
connector-agent-2026-10-08
connector-agent
connector-agent
osv2026-10-08 15:39 UTCMAL-2026-17682link
node-env-resolve-2026-10-08
node-env-resolve
node-env-resolve
osv2026-10-08 14:54 UTCMAL-2026-17681link
bellaxchuu-yarnbails-2026-10-08
@bellaxchuu/yarnbails
@bellaxchuu/yarnbails
osv2026-10-08 14:39 UTCMAL-2026-17680link
bellaxchuu-xbailey-2026-10-08
@bellaxchuu/xbailey
@bellaxchuu/xbailey
osv2026-10-08 14:39 UTCMAL-2026-17678link
bellaxchuu-xbails-2026-10-08
@bellaxchuu/xbails
@bellaxchuu/xbails
osv2026-10-08 14:39 UTCMAL-2026-17679link
bellaxchuu-libsignal-node-2026-10-08
@bellaxchuu/libsignal-node
@bellaxchuu/libsignal-node
osv2026-10-08 14:39 UTCMAL-2026-17676link
bellaxchuu-overbugs-2026-10-08
@bellaxchuu/overbugs
@bellaxchuu/overbugs
osv2026-10-08 14:39 UTCMAL-2026-17677link
bellaxchuu-bails-2026-10-08
@bellaxchuu/bails
@bellaxchuu/bails
osv2026-10-08 14:39 UTCMAL-2026-17675link
dom-utils-lite-2026-10-08
historical OSV index
dom-utils-lite
dom-utils-lite
OSV published 2026-04-14 09:21 UTC
osv2026-10-08 14:38 UTCMAL-2026-2826link
centralogger-2026-10-08
historical OSV index
centralogger
centralogger
OSV published 2026-04-14 09:21 UTC
osv2026-10-08 14:38 UTCMAL-2026-2825link
brat-generator-2026-10-08
brat-generator
brat-generator
osv2026-10-08 13:39 UTCMAL-2026-17674link
xayz-libsignal-node-2026-10-08
@xayz/libsignal-node
@xayz/libsignal-node
osv2026-10-08 13:39 UTCMAL-2026-17672link
bottino-baileys-2026-10-08
bottino-baileys
bottino-baileys
osv2026-10-08 13:39 UTCMAL-2026-17673link
xayz-brat-generator-2026-10-08
@xayz/brat-generator
@xayz/brat-generator
osv2026-10-08 13:39 UTCMAL-2026-17671link
itsmee-aizat-id-baileys2-2026-10-08
@itsmee_aizat.id/baileys2
@itsmee_aizat.id/baileys2
osv2026-10-08 13:39 UTCMAL-2026-17670link
itsmee-aizat-id-baileys-2026-10-08
@itsmee_aizat.id/baileys
@itsmee_aizat.id/baileys
osv2026-10-08 13:39 UTCMAL-2026-17669link
dransay-address-validation-2026-10-08
@dransay/address-validation
@dransay/address-validation
osv2026-10-08 09:08 UTCMAL-2026-17668link
radio-player-theme-2026-10-08
radio-player-theme
radio-player-theme
osv2026-10-08 06:37 UTCMAL-2026-16347link
diezyclutch-baileys-2026-10-08
historical OSV index
diezyclutch-baileys
diezyclutch-baileys
OSV published 2026-08-06 23:56 UTC
osv2026-10-08 06:37 UTCMAL-2026-13482link
tailwind-animatecss-uniform-2026-10-08
tailwind-animatecss-uniform
tailwind-animatecss-uniform
osv2026-10-08 05:22 UTCMAL-2026-17667link
animatecss-tailwind-adapter-2026-10-08
animatecss-tailwind-adapter
animatecss-tailwind-adapter
osv2026-10-08 05:22 UTCMAL-2026-17665link
animatecss-tailwind-bridge-2026-10-08
animatecss-tailwind-bridge
animatecss-tailwind-bridge
osv2026-10-08 05:22 UTCMAL-2026-17666link
wix-ssr-thunderbolt-grid-polyfill-2026-10-08
wix-ssr-thunderbolt-grid-polyfill
wix-ssr-thunderbolt-grid-polyfill
osv2026-10-08 04:07 UTCMAL-2026-17664link
hardhat-pack-2026-10-08
hardhat-pack
hardhat-pack
osv2026-10-08 04:07 UTCMAL-2026-17661link
hardhat-promised-2026-10-08
hardhat-promised
hardhat-promised
osv2026-10-08 04:07 UTCMAL-2026-17662link
mfasolver-2026-10-08
mfasolver
mfasolver
osv2026-10-08 04:07 UTCMAL-2026-17663link
hardhat-deep-2026-10-08
hardhat-deep
hardhat-deep
osv2026-10-08 04:06 UTCMAL-2026-17659link
hardhat-option-2026-10-08
hardhat-option
hardhat-option
osv2026-10-08 04:06 UTCMAL-2026-17660link
css-flow-render-shim-2026-10-08
css-flow-render-shim
css-flow-render-shim
osv2026-10-08 04:06 UTCMAL-2026-17655link
css-reading-display-polyfill-2026-10-08
css-reading-display-polyfill
css-reading-display-polyfill
osv2026-10-08 04:06 UTCMAL-2026-17656link
dotenv-runtime-2026-10-08
dotenv-runtime
dotenv-runtime
osv2026-10-08 04:06 UTCMAL-2026-17657link
hardhat-bits-2026-10-08
hardhat-bits
hardhat-bits
osv2026-10-08 04:06 UTCMAL-2026-17658link
abbishal-poc-as-dependency-2026-10-08
abbishal-poc-as-dependency
abbishal-poc-as-dependency
osv2026-10-08 04:06 UTCMAL-2026-17653link
abbishal-poc2-2026-10-08
abbishal-poc2
abbishal-poc2
osv2026-10-08 04:06 UTCMAL-2026-17654link
chai-as-indexed-2026-10-08
chai-as-indexed
chai-as-indexed
osv2026-10-08 04:06 UTCMAL-2026-16293link
praveenvjpm-color-utils-7210-2026-10-08
@praveenvjpm/color-utils-7210
@praveenvjpm/color-utils-7210
osv2026-10-08 04:06 UTCMAL-2026-17651link
tensorlake-2026-10-08
tensorlake
tensorlake
osv2026-10-08 03:51 UTCMAL-2026-17650link
internallib-v788-2026-10-08
internallib_v788
internallib_v788
osv2026-10-08 02:51 UTCMAL-2026-17649link
function-color-2026-10-07
function-color
function-color
osv2026-10-07 21:35 UTCMAL-2026-17647link
function-flag-2026-10-07
function-flag
function-flag
osv2026-10-07 21:35 UTCMAL-2026-17648link
tailwindcss-animatecss-keyframes-2026-10-07
tailwindcss-animatecss-keyframes
tailwindcss-animatecss-keyframes
osv2026-10-07 16:48 UTCMAL-2026-17646link
vectorizers-2026-10-07
vectorizers
vectorizers
osv2026-10-07 13:42 UTCMAL-2026-17645link
random-certs-2026-10-07
random-certs
random-certs
osv2026-10-07 13:42 UTCMAL-2026-17644link
waie-crash-baileys-2026-10-07
waie-crash-baileys
waie-crash-baileys
osv2026-10-07 10:42 UTCMAL-2026-17643link
browser-metrics-plugin-contrib-2026-10-07
browser-metrics-plugin.contrib
browser-metrics-plugin.contrib
osv2026-10-07 06:56 UTCMAL-2026-17640link
troubleshooting-2026-10-07
troubleshooting
troubleshooting
osv2026-10-07 06:56 UTCMAL-2026-17641link
personio-pipeline-projen-2026-10-07
personio-pipeline-projen
personio-pipeline-projen
osv2026-10-07 06:56 UTCMAL-2026-17642link
store-webpack-config-2026-10-07
historical OSV index
store-webpack-config
store-webpack-config
OSV published 2025-10-30 17:38 UTC
osv2026-10-07 06:41 UTCMAL-2025-49241link
windows-registry-editor-2026-10-07
historical OSV index
windows_registry_editor
windows_registry_editor
OSV published 2025-10-30 17:38 UTC
osv2026-10-07 06:41 UTCMAL-2025-49244link
eslint-config-support-2026-10-07
historical OSV index
eslint-config-support
eslint-config-support
OSV published 2022-06-20 20:09 UTC
osv2026-10-07 06:41 UTCMAL-2022-2806link
lore-cs-2026-10-06
lore-cs
lore-cs
osv2026-10-06 21:56 UTCMAL-2026-17639link
openapi-python-client-2026-10-06
openapi-python-client
openapi-python-client
osv2026-10-06 20:37 UTCGHSA-5293-mq8x-g3xjlink
css-jptvix-polyfill-2026-10-06
css-jptvix-polyfill
css-jptvix-polyfill
osv2026-10-06 20:26 UTCMAL-2026-17638link
captchetat-angularv8-2026-10-06
captchetat-angularv8
captchetat-angularv8
osv2026-10-06 12:54 UTCMAL-2026-17637link
hardhat-init-2026-10-06
hardhat-init
hardhat-init
osv2026-10-06 04:52 UTCMAL-2026-17636link
internallib-v23-2026-10-06
internallib_v23
internallib_v23
osv2026-10-06 04:37 UTCMAL-2026-17632link
internallib-v30-2026-10-06
internallib_v30
internallib_v30
osv2026-10-06 04:37 UTCMAL-2026-17633link
internallib-v86-2026-10-06
internallib_v86
internallib_v86
osv2026-10-06 04:37 UTCMAL-2026-17634link
internallib-v875-2026-10-06
internallib_v875
internallib_v875
osv2026-10-06 04:37 UTCMAL-2026-17635link
pinecone-experience-messages-2026-10-06
@pinecone-experience/messages
@pinecone-experience/messages
osv2026-10-06 04:37 UTCMAL-2026-17631link
langflow-2026-10-06
langflow
langflow
osv2026-10-06 02:36 UTCGHSA-qqw9-2vj9-hx7rlink
virgil-cli-2026-10-05
virgil-cli
virgil-cli
osv2026-10-05 23:21 UTCMAL-2026-17630link
zencleaner-2026-10-05
zencleaner
zencleaner
osv2026-10-05 23:06 UTCMAL-2026-17629link
erc7730-hackathon-2026-10-05
erc7730-hackathon
erc7730-hackathon
osv2026-10-05 21:35 UTCMAL-2026-17628link
node-env-resolver-dotenvx-2026-10-05
node-env-resolver-dotenvx
node-env-resolver-dotenvx
npm_deprecation2026-10-05 20:40 UTC—link
documenclient-2026-10-05
documenclient
documenclient
osv2026-10-05 18:50 UTCMAL-2026-17624link
serpacksven-2026-10-05
serpacksven
serpacksven
osv2026-10-05 18:50 UTCMAL-2026-17625link
serpacksven1-2026-10-05
serpacksven1
serpacksven1
osv2026-10-05 18:50 UTCMAL-2026-17626link
serpacksven2-2026-10-05
serpacksven2
serpacksven2
osv2026-10-05 18:50 UTCMAL-2026-17627link
css-yhpodl-polyfill-2026-10-05
css-yhpodl-polyfill
css-yhpodl-polyfill
osv2026-10-05 16:34 UTCMAL-2026-17578link
with-cte-2026-10-05
with-cte
with-cte
osv2026-10-05 16:34 UTCMAL-2026-17579link
commit lag

ledger_commit_at is missing or more than 24 hours after attestd_detected_at. Same incident class as the 24h-verifiable rows. Detection time is still first_seen / attestd_detected_at, not commit time.

Showing 100 of 15237. Show all in this section · Full ledger.json

IncidentSourceRegisteredOSVCitation
xayz-baileys-2026-09-30
@xayz/baileys
@xayz/baileys
osv2026-09-30 10:33 UTCMAL-2026-17384link
lendxntaa-baileys-2026-09-30
@lendxntaa/baileys
@lendxntaa/baileys
osv2026-09-30 10:33 UTCMAL-2026-17365link
kelvdra-baileys-2026-09-30
@kelvdra/baileys
@kelvdra/baileys
osv2026-09-30 10:33 UTCMAL-2026-17362link
kxa-xbails-2026-09-30
@kxa/xbails
@kxa/xbails
osv2026-09-30 10:33 UTCMAL-2026-17363link
japofc-baileys-2026-09-30
@japofc/baileys
@japofc/baileys
osv2026-09-30 10:32 UTCMAL-2026-17359link
bottino-baileys-2026-09-30
@bottino/baileys
@bottino/baileys
osv2026-09-30 10:32 UTCMAL-2026-17351link
badzz88-baileys-2026-09-30
@badzz88/baileys
@badzz88/baileys
osv2026-09-30 10:32 UTCMAL-2026-17341link
imgbundle-2026-09-30
imgbundle
imgbundle
osv2026-09-30 05:30 UTCMAL-2026-17330link
cdn-img-fetch-2026-09-30
cdn-img-fetch
cdn-img-fetch
osv2026-09-30 05:00 UTCMAL-2026-17320link
tailwind-forms-kit-2026-09-30
tailwind-forms-kit
tailwind-forms-kit
osv2026-09-30 01:44 UTCMAL-2026-17310link
akapaki-baileys-2026-09-29
@akapaki/baileys
@akapaki/baileys
osv2026-09-29 21:26 UTCMAL-2026-17285link
exprrdd-2026-09-29
exprrdd
exprrdd
osv2026-09-29 14:53 UTCMAL-2026-17244link
exptred-2026-09-29
exptred
exptred
osv2026-09-29 14:53 UTCMAL-2026-17246link
exptredd-2026-09-29
exptredd
exptredd
osv2026-09-29 14:53 UTCMAL-2026-17247link
xeprews-2026-09-29
xeprews
xeprews
osv2026-09-29 14:53 UTCMAL-2026-17248link
exprdd-2026-09-29
exprdd
exprdd
osv2026-09-29 14:53 UTCMAL-2026-17241link
express-javascript-2026-09-29
express-javascript
express-javascript
osv2026-09-29 14:53 UTCMAL-2026-17242link
express-nodejs-2026-09-29
express-nodejs
express-nodejs
osv2026-09-29 14:53 UTCMAL-2026-17243link
items-validator-2026-09-28
items-validator
items-validator
osv2026-09-28 22:34 UTCMAL-2026-17235link
figlet-chalk-render-2026-09-28
figlet-chalk-render
figlet-chalk-render
osv2026-09-28 18:33 UTCMAL-2026-17226link
digift-cli-2026-09-25
@digift/cli
@digift/cli
osv2026-09-25 18:59 UTCMAL-2026-17187link
alphaspace-core-2026-09-25
@alphaspace/core
@alphaspace/core
osv2026-09-25 02:58 UTCMAL-2026-17169link
tailwind-contact-forms-2026-09-23
tailwind-contact-forms
tailwind-contact-forms
osv2026-09-23 14:32 UTCMAL-2026-15925link
com-apple-unityplugin-storekit-2026-09-23
com.apple.unityplugin.storekit
com.apple.unityplugin.storekit
osv2026-09-23 14:32 UTCMAL-2026-16477link
pullgetsage-2026-09-21
pullgetsage
pullgetsage
osv2026-09-21 20:30 UTCMAL-2026-16366link
baanx-blockchain-config-2026-09-21
@baanx/blockchain-config
@baanx/blockchain-config
osv2026-09-21 17:42 UTCMAL-2026-16352link
woodpecker-web-shared-components-2026-09-21
@woodpecker-web-shared/components
@woodpecker-web-shared/components
osv2026-09-21 17:42 UTCMAL-2026-16354link
baanx-abis-2026-09-21
@baanx/abis
@baanx/abis
osv2026-09-21 17:42 UTCMAL-2026-16351link
hardhat-devkit-2026-09-21
hardhat-devkit
hardhat-devkit
osv2026-09-21 16:57 UTCMAL-2026-16349link
feed-widget-helper-2026-09-21
feed-widget-helper
feed-widget-helper
osv2026-09-21 03:54 UTCMAL-2026-16332link
homestack-cheer-2026-09-21
homestack-cheer
homestack-cheer
osv2026-09-21 03:54 UTCMAL-2026-16333link
baanx-solana-lib-2026-09-21
@baanx/solana-lib
@baanx/solana-lib
osv2026-09-21 03:39 UTCMAL-2026-16300link
pwaplatform-module-sso-integration-2026-09-20
@pwaplatform/module-sso-integration
@pwaplatform/module-sso-integration
osv2026-09-20 11:30 UTCMAL-2026-16299link
insiderintelligence-googleadmanager-2026-09-18
@insiderintelligence/googleadmanager
@insiderintelligence/googleadmanager
osv2026-09-18 20:54 UTCMAL-2026-16290link
lekzo-baileys-2026-09-18
@lekzo/baileys
@lekzo/baileys
osv2026-09-18 02:28 UTCMAL-2026-16276link
tailwind-forms-styles-2026-09-15
tailwind-forms-styles
tailwind-forms-styles
osv2026-09-15 16:26 UTCMAL-2026-16195link
id79-client-2026-09-15
id79-client
id79-client
osv2026-09-15 06:36 UTCMAL-2026-16173link
biz44-id79-client-2026-09-15
@biz44/id79-client
@biz44/id79-client
osv2026-09-15 06:36 UTCMAL-2026-16168link
biz44-id95-client-2026-09-15
@biz44/id95-client
@biz44/id95-client
osv2026-09-15 06:36 UTCMAL-2026-16169link
biz44-id99-client-2026-09-15
@biz44/id99-client
@biz44/id99-client
osv2026-09-15 06:36 UTCMAL-2026-16170link
biz44-process-runtime-utils-2026-09-15
@biz44/process-runtime-utils
@biz44/process-runtime-utils
osv2026-09-15 06:36 UTCMAL-2026-16171link
biz44-runtime-utils-2026-09-15
@biz44/runtime-utils
@biz44/runtime-utils
osv2026-09-15 06:36 UTCMAL-2026-16172link
biz44-id10-client-2026-09-15
@biz44/id10-client
@biz44/id10-client
osv2026-09-15 06:36 UTCMAL-2026-16165link
biz44-id12-client-2026-09-15
@biz44/id12-client
@biz44/id12-client
osv2026-09-15 06:36 UTCMAL-2026-16166link
biz44-id44-client-2026-09-15
@biz44/id44-client
@biz44/id44-client
osv2026-09-15 06:36 UTCMAL-2026-16167link
yggbrasil-api-2026-09-14
@yggbrasil/api
@yggbrasil/api
osv2026-09-14 22:34 UTCMAL-2026-16163link
n8n-nodes-sysdiag2-2026-09-14
n8n-nodes-sysdiag2
n8n-nodes-sysdiag2
osv2026-09-14 22:19 UTCMAL-2026-16162link
lpulogin-2026-09-14
lpulogin
lpulogin
osv2026-09-14 20:33 UTCMAL-2026-16160link
meraz-project-tracker-2026-09-14
meraz-project-tracker
meraz-project-tracker
osv2026-09-14 20:33 UTCMAL-2026-16161link
afhmxiewpsf-2026-09-14
afhmxiewpsf
afhmxiewpsf
osv2026-09-14 20:33 UTCMAL-2026-16158link
dilxztech-2026-09-14
dilxztech
dilxztech
osv2026-09-14 20:33 UTCMAL-2026-16159link
merazmz-project-tracker-2026-09-14
@merazmz/project-tracker
@merazmz/project-tracker
osv2026-09-14 20:33 UTCMAL-2026-16157link
get-power-2026-09-14
get-power
get-power
osv2026-09-14 19:48 UTCMAL-2026-16156link
pino-ulid-2026-09-14
pino-ulid
pino-ulid
osv2026-09-14 18:32 UTCMAL-2026-16154link
ultra-ws-2026-09-14
ultra-ws
ultra-ws
osv2026-09-14 18:32 UTCMAL-2026-16155link
sql-limit-enforcer-2026-09-14
sql-limit-enforcer
sql-limit-enforcer
osv2026-09-14 17:31 UTCMAL-2026-16151link
strapi-plugin-os-info-meeb322k-2026-09-14
strapi-plugin-os-info-meeb322k
strapi-plugin-os-info-meeb322k
osv2026-09-14 17:31 UTCMAL-2026-16152link
web-main-2026-09-14
web-main
web-main
osv2026-09-14 17:31 UTCMAL-2026-16153link
n8n-nodes-sysdiag-2026-09-14
n8n-nodes-sysdiag
n8n-nodes-sysdiag
osv2026-09-14 17:31 UTCMAL-2026-16147link
noblox-asset-js-2026-09-14
noblox-asset.js
noblox-asset.js
osv2026-09-14 17:31 UTCMAL-2026-16148link
os-info-meeb322k-2026-09-14
os-info-meeb322k
os-info-meeb322k
osv2026-09-14 17:31 UTCMAL-2026-16149link
postgreesqlhelper-2026-09-14
postgreesqlhelper
postgreesqlhelper
osv2026-09-14 17:31 UTCMAL-2026-16150link
discord-mfa-solver-2026-09-14
discord-mfa-solver
discord-mfa-solver
osv2026-09-14 16:46 UTCMAL-2026-16100link
aiwfm-communitywfm-scripts-api-2026-09-14
@aiwfm/communitywfm.scripts.api
@aiwfm/communitywfm.scripts.api
osv2026-09-14 16:46 UTCMAL-2026-16146link
app-rrhh-2026-09-14
app-rrhh
app-rrhh
osv2026-09-14 16:31 UTCMAL-2026-16144link
concierge-sdk-2026-09-14
concierge-sdk
concierge-sdk
osv2026-09-14 16:31 UTCMAL-2026-16145link
chroma-client-2026-09-13
chroma-client
chroma-client
osv2026-09-13 17:27 UTCMAL-2026-16143link
mountly-tailwind-2026-09-13
mountly-tailwind
mountly-tailwind
npm_deprecation2026-09-13 08:16 UTC—link
python-fork-2026-09-12
python-fork
python-fork
osv2026-09-12 10:23 UTCMAL-2026-16142link
medialane-ui-2026-09-12
@medialane/ui
@medialane/ui
npm_deprecation2026-09-12 01:43 UTC—link
platform-telemetry-client-2026-09-12
platform-telemetry-client
platform-telemetry-client
osv2026-09-12 00:51 UTCMAL-2026-16141link
tailwindcss-contact-forms-2026-09-11
tailwindcss-contact-forms
tailwindcss-contact-forms
osv2026-09-11 23:20 UTCMAL-2026-16124link
tracker-cloudflare-2026-09-11
tracker-cloudflare
tracker-cloudflare
osv2026-09-11 23:20 UTCMAL-2026-16140link
tailwind-form-kit-2026-09-11
tailwind-form-kit
tailwind-form-kit
osv2026-09-11 23:20 UTCMAL-2026-16139link
cr-bot-common-2026-09-11
cr-bot-common
cr-bot-common
osv2026-09-11 23:20 UTCMAL-2026-16137link
greensaver-2026-09-11
greensaver
greensaver
osv2026-09-11 23:20 UTCMAL-2026-16138link
ollamaa-2026-09-11
ollamaa
ollamaa
osv2026-09-11 21:18 UTCMAL-2026-16134link
openaii-2026-09-11
openaii
openaii
osv2026-09-11 21:18 UTCMAL-2026-16135link
transfomers-2026-09-11
transfomers
transfomers
osv2026-09-11 21:18 UTCMAL-2026-16136link
langgrap-2026-09-11
langgrap
langgrap
osv2026-09-11 21:18 UTCMAL-2026-16133link
zalastax-nolb-ngx-s-2026-09-08
historical OSV index
@zalastax/nolb-ngx-s
@zalastax/nolb-ngx-s
OSV published 2025-08-14 18:52 UTC
osv2026-09-08 01:59 UTCMAL-2025-12547link
citra-bubursumsum87-riris-2026-09-08
historical OSV index
citra-bubursumsum87-riris
citra-bubursumsum87-riris
OSV published 2025-11-11 20:46 UTC
osv2026-09-08 01:59 UTCMAL-2025-125469link
citra-asinan27-riris-2026-09-08
historical OSV index
citra-asinan27-riris
citra-asinan27-riris
OSV published 2025-11-11 20:46 UTC
osv2026-09-08 01:59 UTCMAL-2025-125468link
circular-haddock-z3n-2026-09-08
historical OSV index
circular_haddock_z3n
circular_haddock_z3n
OSV published 2025-11-11 20:46 UTC
osv2026-09-08 01:59 UTCMAL-2025-125467link
circular-boar-z3n-2026-09-08
historical OSV index
circular_boar_z3n
circular_boar_z3n
OSV published 2025-11-11 20:46 UTC
osv2026-09-08 01:59 UTCMAL-2025-125466link
cindy-ubi99-riris-2026-09-08
historical OSV index
cindy-ubi99-riris
cindy-ubi99-riris
OSV published 2025-11-11 20:46 UTC
osv2026-09-08 01:59 UTCMAL-2025-125465link
cindy-tumis38-riris-2026-09-08
historical OSV index
cindy-tumis38-riris
cindy-tumis38-riris
OSV published 2025-11-11 20:46 UTC
osv2026-09-08 01:59 UTCMAL-2025-125464link
cindy-tongseng9-riris-2026-09-08
historical OSV index
cindy-tongseng9-riris
cindy-tongseng9-riris
OSV published 2025-11-11 20:46 UTC
osv2026-09-08 01:59 UTCMAL-2025-125463link
cindy-tempe93-riris-2026-09-08
historical OSV index
cindy-tempe93-riris
cindy-tempe93-riris
OSV published 2025-11-11 20:46 UTC
osv2026-09-08 01:59 UTCMAL-2025-125462link
cindy-telurtahu85-riris-2026-09-08
historical OSV index
cindy-telurtahu85-riris
cindy-telurtahu85-riris
OSV published 2025-11-11 20:46 UTC
osv2026-09-08 01:59 UTCMAL-2025-125461link
cindy-telur50-riris-2026-09-08
historical OSV index
cindy-telur50-riris
cindy-telur50-riris
OSV published 2025-11-11 20:46 UTC
osv2026-09-08 01:59 UTCMAL-2025-125460link
zalastax-nolb-ngx-r-2026-09-08
historical OSV index
@zalastax/nolb-ngx-r
@zalastax/nolb-ngx-r
OSV published 2025-08-14 18:52 UTC
osv2026-09-08 01:59 UTCMAL-2025-12546link
cindy-tahu87-riris-2026-09-08
historical OSV index
cindy-tahu87-riris
cindy-tahu87-riris
OSV published 2025-11-11 20:46 UTC
osv2026-09-08 01:59 UTCMAL-2025-125459link
cindy-sroto54-riris-2026-09-08
historical OSV index
cindy-sroto54-riris
cindy-sroto54-riris
OSV published 2025-11-11 20:46 UTC
osv2026-09-08 01:59 UTCMAL-2025-125458link
cindy-soto66-riris-2026-09-08
historical OSV index
cindy-soto66-riris
cindy-soto66-riris
OSV published 2025-11-11 20:46 UTC
osv2026-09-08 01:59 UTCMAL-2025-125457link
cindy-semur24-riris-2026-09-08
historical OSV index
cindy-semur24-riris
cindy-semur24-riris
OSV published 2025-11-11 20:46 UTC
osv2026-09-08 01:59 UTCMAL-2025-125456link
cindy-ruwet97-riris-2026-09-08
historical OSV index
cindy-ruwet97-riris
cindy-ruwet97-riris
OSV published 2025-11-11 20:46 UTC
osv2026-09-08 01:59 UTCMAL-2025-125455link
cindy-rangginang86-riris-2026-09-08
historical OSV index
cindy-rangginang86-riris
cindy-rangginang86-riris
OSV published 2025-11-11 20:46 UTC
osv2026-09-08 01:59 UTCMAL-2025-125454link
cindy-ragi78-riris-2026-09-08
historical OSV index
cindy-ragi78-riris
cindy-ragi78-riris
OSV published 2025-11-11 20:46 UTC
osv2026-09-08 01:59 UTCMAL-2025-125453link
cindy-pisang38-riris-2026-09-08
historical OSV index
cindy-pisang38-riris
cindy-pisang38-riris
OSV published 2025-11-11 20:46 UTC
osv2026-09-08 01:59 UTCMAL-2025-125452link

historical OSV index = source_published_at, or the OSV published date when that field is absent, is 30 days or more before attestd_detected_at. Not a separate incident class. Rows without those dates are unlabeled.

methodology

How entries are defined

Registered is attestd_detected_at: the earliest first_seen, when the API began returning compromised: true. It is not GitHub commit time. 24h-verifiable means the ledger commit landed within 24 hours of that registration. Public git history confirms the lag bound. It does not make commit time equal detection time.

historical OSV index is a per-row label, not a separate incident class. It applies when source_published_at, or the OSV published date if that field is absent, is 30 days or more before attestd_detected_at. It is independent of whether ledger_commit_at landed within 24 hours.

OSV is a citation to the linked advisory when Attestd has one. Most live entries are registered from the OSV ingestion pipeline. An OSV link means Attestd indexed that advisory; it is not a claim that Attestd published first.

Citation is the primary advisory URL, a press URL when recorded, or the OSV page.

Detection source discloses how the incident entered Attestd. osv is the OSV MAL pipeline. registry is curated registry YAML. npm_deprecation and pypi_yank are registry-native security signals. news_monitor and agent are corroborated news or agent promotions.

Commits are public on GitHub. Manual historical sync may use GPG signing when configured. Live auto-push uses the GitHub Git Data API so incident files and ledger.json update together. Verify history with git log in the repository.

related