# What share of the vulnerabilities CISA adds to its Known Exploited Vulnerabilities catalog from 1 January to 31 March 2027 will have a fix deadline of three days or less?

> As of 13 November 2026, Prolepsia’s median forecast is 77, with an 80% chance it falls between 64.5 and 87.

- Forecaster: Prolepsia (https://prolepsia.com)
- Forecast made: 2026-11-13
- Question type: Number
- Resolves: 2027-04-02
- Status: Open
- Forecasts: 1
- Page: https://prolepsia.com/forecasts/variance-lattice-almanac/what-share-of-the-vulnerabilities-cisa-adds-to-its-known-exploited-vulnerabiliti
- Open in the app: https://app.prolepsia.com/library/e5985247-601d-4714-bfe6-46f04c055f61

## Forecast

- 10th percentile: 64.5
- Median: 77
- 90th percentile: 87

## How Prolepsia reasons

## TL;DR
The forecast for new KEV entries added during 1 January–31 March 2027 has a median of 77%, a 10th percentile of 64.5%, and a 90th percentile of 87% receiving deadlines of three days or less. Plan around short deadlines remaining the dominant category, with longer deadlines still part of the system. The decisive distinction is between a durable risk-based policy and an uninterrupted march toward universal three-day treatment.

## Context
CISA’s Known Exploited Vulnerabilities catalog records vulnerabilities with evidence of exploitation and assigns remediation deadlines. The question measures the calendar-day difference between each entry’s `dateAdded` and `dueDate`, not whether agencies actually fix the vulnerability. Resolution uses distinct CVEs added during the target quarter that remain in the catalog at 23:59 US Eastern on 1 April 2027, including any deadline edits made by that snapshot. The designated source is the [official catalog](https://www.cisa.gov/sites/default/files/feeds/known_exploited_vulnerabilities.json), with the [official repository](https://github.com/cisagov/kev-data) as the fallback.

The forecast date is 13 November 2026. The latest catalog edition established at that point was version 2026.11.10, released on 10 November 2026; a download on the forecast date therefore did not represent three additional days of catalog updates ([catalog mirror](https://raw.githubusercontent.com/cisagov/kev-data/main/known_exploited_vulnerabilities.json)). The central question is whether the recent pattern persists through the next quarter, not whether the older deadline regime returns by default.

## Evidence
The historical backbone is a sharp change in deadline assignment, followed by a less directional recent pattern. The full quarterly history shows why neither the older baseline nor a straight-line extrapolation is sufficient. These are counts of entries with deadlines of three days or less, grouped by addition date in the November catalog snapshot ([catalog](https://www.cisa.gov/sites/default/files/feeds/known_exploited_vulnerabilities.json)).

| Addition period | Deadline ≤3 days | All additions |
|---|---:|---:|
| 2025 Q1 | 0 | 73 |
| 2025 Q2 | 0 | 59 |
| 2025 Q3 | 7 | 51 |
| 2025 Q4 | 0 | 62 |
| 2026 Q1 | 13 | 71 |
| 2026 Q2 | 32 | 75 |
| 2026 Q3 | 78 | 100 |
| 1 October–10 November 2026 | 36 | 49 |

The early observations establish how unusual short deadlines once were. The later observations establish that they became the dominant category. But the partial fourth quarter does not show another clear increase over the third quarter. The data supports a changed regime more strongly than it supports continued acceleration.

There is a policy reason to give the newer observations greater relevance. **BOD 26-04 superseded BOD 22-01**, so the question’s description of the governing framework is outdated. Issued on 10 June 2026, the newer directive prioritizes remediation using exposure, KEV status, exploitation automatability, and technical impact ([directive](https://www.cisa.gov/news-events/directives/bod-26-04-prioritizing-security-updates-based-risk)). This is a structural explanation for the shift, rather than merely a run of unusually urgent alerts.

The finer-grained history also matters. The table below gives the complete monthly addition-volume sequence available from May through the observed part of November, alongside the short-deadline counts established for the recent months. A dash means no full-month qualifying count is stated, not zero ([catalog records](https://raw.githubusercontent.com/cisagov/kev-data/main/known_exploited_vulnerabilities.csv)).

| Addition period, 2026 | Deadline ≤3 days | All additions |
|---|---:|---:|
| May | — | 21 |
| June | — | 23 |
| July | 22 | 26 |
| August | 22 | 31 |
| September | 34 | 43 |
| October | 29 | 36 |
| 1–10 November | 7 | 13 |

For the post-directive portion of June, 11–30 June, the count was 12 of 13. Across 11 June–10 November, it was 126 of 162, or 77.78%. Restricting the window to July–10 November gives 114 of 149, or 76.51% ([catalog mirror](https://raw.githubusercontent.com/cisagov/kev-data/main/known_exploited_vulnerabilities.json)). These overlapping historical windows support the forecast’s central location without establishing a fixed future rate.

The recent weakness deserves attention. From 20 October–10 November, 18 of 31 additions had short deadlines, compared with 108 of 131 from 11 June–19 October. That is a real counterpoint to the continuity case. It is not proof of a lasting reversal: the split was identified after examining the observations, the later window is small, and related CVEs can arrive in batches. The forecast retains room for a lower quarterly share rather than treating the dip as a new settled baseline.

Longer deadlines are not simply leftovers from the old system. On 10 November, two Barracuda additions received three days, while the other four additions received fourteen days ([dated catalog update](https://github.com/cisagov/kev-data/commit/6b8d7ba579672e4f84e2b1f255724aa97c8455b4)). CISA’s implementation guidance connects catalog due dates to its risk table and available exposure evidence. It also distinguishes catalog deadlines from actual asset-level requirements ([implementation guidance](https://www.cisa.gov/news-events/directives/bod-26-04-implementation-guidance-prioritizing-security-updates-based-risk)). A continuing mixture is therefore consistent with the policy working as intended.

The decision table makes that distinction concrete. Four of its eight KEV combinations receive three-day timelines and four receive fourteen-day timelines ([risk decision table](https://raw.githubusercontent.com/CERTCC/SSVC/675fda1e1c7c7d97763ff57ce549183da61f2fba/data/json/decision_tables/cisa/cisa_bod_26_04_1_0_0.json)). Those rows are not equally common populations. Their existence explains why longer deadlines remain legitimate; their frequency among actual additions determines the measured share.

The December milestone supports policy persistence, but not an automatic increase. FedRAMP’s 16 June notice requires agencies to begin evaluating and remediating under the new timelines by 7 December 2026 and describes a corrective-action grace period through 7 March 2027 ([FedRAMP notice](https://www.fedramp.gov/notices/0014/)). Its 10 November notice allows case-by-case conditional extensions beyond that March date ([follow-up notice](https://www.fedramp.gov/notices/0017/)). These are implementation and enforcement provisions. They do not establish that CISA will shorten or lengthen the catalog’s date fields.

## What's non-obvious
The obvious reading is that the government has been tightening deadlines and will keep doing so. The better reading is that a risk-based system has replaced an older pattern, and the recent share reflects the vulnerabilities flowing through that system. **A change in the share need not mean a change in government urgency.** Different exposure assessments, technical impacts, or exploitation characteristics can change the result under the same rules. Conversely, stronger implementation pressure can leave the catalog share unchanged.

A second distinction is between vulnerability counts and independent events. Several entries can concern related products or exploitation campaigns and receive similar deadlines together. A busy addition day can therefore move the share more than an independent-entry interpretation suggests. Quarterly volume matters too: the same batch has more influence in a smaller quarter. These features explain why a recent historical share close to the forecast median does not justify a narrow prediction interval.

## Uncertainties
The largest unknown is the target quarter’s vulnerability composition. No January–March entries are observed at the forecast date. A quarter concentrated in high-risk, exposed, readily exploitable vulnerabilities would push the short-deadline share upward. More entries qualifying for longer treatment would push it downward. Neither requires a formal policy reversal.

The recent decline is the clearest downside signal, but its persistence is unverified. More complete November and December cohorts, including entry-level deadline categories and product groupings, would help distinguish a temporary batch effect from a sustained change. An explicit announcement changing catalog date assignment would be stronger evidence of a policy discontinuity than an agency implementation milestone alone.

Final-snapshot revisions create a separate uncertainty. The historical counts describe surviving entries and their recorded deadlines in the November edition; they do not reconstruct every deadline as originally published. There is no comprehensive historical revision series here establishing how often dates are shortened, extended, or entries removed. A dated record of those changes would close that gap. CISA’s guidance says timeline changes should be infrequent and evidence-supported, but that is a policy statement, not a measured revision frequency ([guidance](https://www.cisa.gov/news-events/directives/bod-26-04-implementation-guidance-prioritizing-security-updates-based-risk)).

The 10th and 90th percentiles describe uncertainty about the realized quarterly share, not hard bounds and not agency compliance. The forecast is conditional on positive additions; if the quarter has none, the question is annulled. The main conclusion remains continuity of a dominant short-deadline category, with meaningful variation around it.

## Forecast history

| Date | Forecast |
|---|---|
| 2026-11-13 | Median 77, 80% range 64.5 to 87 |

## Sources Prolepsia read

- [official catalog](https://www.cisa.gov/sites/default/files/feeds/known_exploited_vulnerabilities.json)
- [GitHub - cisagov/kev-data: Mirror of cisa.gov/kev data files · GitHub](https://github.com/cisagov/kev-data)
- [catalog mirror](https://raw.githubusercontent.com/cisagov/kev-data/main/known_exploited_vulnerabilities.json)
- [FedRAMP Response to CISA BOD 26-04 (Prioritizing Security Updates Based on Risk)](https://www.fedramp.gov/notices/0014)
- [CISA Releases Binding Operational Directive 26-04 on Prioritizing Security Updates Based on Risk](https://content.govdelivery.com/accounts/USDHSCISA/bulletins/41b445a)
- [BOD 26-04: Prioritizing Security Updates Based on Risk \| CISA](https://www.dejavu.org/cgi-bin/get.cgi?url=https%3A%2F%2Fwww.cisa.gov%2Fnews-events%2Fdirectives%2Fbod-26-04-prioritizing-security-updates-based-risk&ver=93)

## The question

CISA's Known Exploited Vulnerabilities (KEV) catalog lists vulnerabilities with evidence of exploitation. Each entry has a dateAdded and a dueDate, the deadline by which US federal civilian agencies must fix it under Binding Operational Directive 22-01. For years the usual deadline was 21 days. From July 2025 CISA began setting deadlines of three days or less, and by late 2026 most new entries had one. Share of new entries with a deadline of three days or less, from the official catalog downloaded on 13 November 2026: 2025 Q1 0 of 73, Q2 0 of 59, Q3 7 of 51, Q4 0 of 62; 2026 Q1 13 of 71 (18.3%), Q2 32 of 75 (42.7%), Q3 78 of 100 (78.0%), and 1 October to 10 November 36 of 49 (73.5%). The other new entries in 2026 mostly had 14 or 21 days. This question measures how hard the government pushes for fixing exploited vulnerabilities within days.

### How it resolves

The question resolves to 100 times the number of entries with dateAdded from 2027-01-01 to 2027-03-31 inclusive whose dueDate is three calendar days or less after their dateAdded, divided by the number of all entries with dateAdded in that range. Entries are distinct CVE identifiers in CISA's official catalog (https://www.cisa.gov/sites/default/files/feeds/known_exploited_vulnerabilities.json), as published at 23:59 US Eastern Time on 1 April 2027. If that version is unavailable, use the cisagov/kev-data GitHub repository at the same time. Use the dueDate in that version, even if CISA changed it after the entry was added. If CISA adds no entries in the period, the question is annulled.

### Fine print

Days are calendar days computed from the two date fields as written. An entry removed before 1 April 2027 does not count. The question does not ask whether agencies met the deadlines.

---

Prolepsia’s forecasting engine made this forecast: it researched the question, weighed the evidence and wrote the report on this page. A probability is not a promise; this one changes as the evidence does.
