8000 [pull] main from CPAtoCybersecurity:main by pull[bot] · Pull Request #24 · ericrihm/csf_profile · GitHub
[go: up one dir, main page]

Skip to content

[pull] main from CPAtoCybersecurity:main - #24

Merged
pull[bot] merged 16 commits into
ericrihm:mainfrom
CPAtoCybersecurity:main
Jul 19, 2026
Merged

pull[bot] merged 16 commits into
ericrihm:mainfrom
CPAtoCybersecurity:main

Conversation

@pull
@pull pull Bot commented Jul 19, 2026 •
Copy link
Copy Markdown

See Commits and Changes for more details.


Created by pull[bot] (v2.0.0-alpha.4)

Can you help keep this open source service alive? 💖 Please sponsor : )

…t 5-point CMMI scale option (#277)

- New ScoreSelect component: whole-number + quarter-fraction (.00/.25/.50/.75)
  dropdowns with fraction locked at scale max; replaces the integer-only
  selects in Assessments and the half-point selects in Controls.
- New-assessment wizard step 1 asks 10-point (default, Simply Cyber Academy)
  vs 5-point CMMI-style scale (NIST CSF Maturity Toolkit); scoringScale is
  stored per assessment and locked at creation.
- assessmentsStore schema v10: stamps scoringScale 10 on existing
  assessments; scores are never rescaled (byte-equality tested).
- New scoringScale util is the single source of truth: compose/decompose/
  snap, proportional score bands (10-scale thresholds unchanged), CMMI
  level reference.
- Scale-aware consumers: score badge, audit report (CMMI Appendix B,
  dynamic scale max, quarter-aware formatting), executive PDF (fixes
  pre-existing "/ 7.0" on 10-point data), AI gap severity, Dashboard chart
  domains (fixes pre-existing [0,7] trend axis) and priority bands,
  Scoring Legend CMMI table, AI score suggestion decimal parse + snap.
- CSV exports gain a Scoring Scale column; import round-trips it
  (header-keyed; old CSVs default to 10). JSON restore migrates pre-v10
  exports; pack imports stamp scale 10 (packFormat 1 contract).
- 40 new tests; suite 271/271.

Closes #277
The "Set up org profile" button appeared to do nothing. Root cause was not the
button (its onClick, state, and render gate are all correct, and the wizard does
mount) — the app has no Tailwind build and src/index.css never defined `.fixed`
or `.inset-0`, so every `fixed inset-0` overlay computed `position: static` and
rendered in document flow. The org-profile wizard opened ~1300px below the
viewport, behind the (also in-flow) New Assessment modal, so nothing appeared.

CSS-only fix in src/index.css:
- .fixed { position: fixed; }  .inset-0 { top/right/bottom/left: 0 }  .bottom-0
- .max-w-lg { max-width: 32rem } so dialog cards are constrained, not full-width
- modal-backdrop dim: .bg-black reads --tw-bg-opacity; .bg-opacity-50 sets it to
  0.5 (color-agnostic, never forces black on other bgs); .bg-black/50, .bg-black/10

This repairs every `fixed inset-0` overlay app-wide (New Assessment modal,
Settings dialogs, ExportPasswordDialog, dropdowns, toolbars), not just the org
wizard. Verified live: the wizard now opens as a centered dialog with a dark
backdrop on top; New Assessment modal, DropdownPortal, and base views render
correctly with no lingering overlay. Added src/index.overlays.test.js guarding
the utilities. No component JS changed.
…verlays

Fix #279: define .fixed/.inset-0 so modal overlays position correctly
… at 90vh

The New Assessment modal card is `max-h-[90vh] flex flex-col` with a fixed
header (name field), a `flex-1 overflow-auto` body (the requirement/scope
picker), and a fixed footer (Cancel/Skip/Next). The overlay centers the card
with `items-center`, so when the card is taller than the viewport its header and
footer render off-screen and neither the body nor the page scrolls to reach
them — you can't name the assessment or press Next.

Same inert-utility root cause as #279: the app has no Tailwind build, and
src/index.css (hand-written utility subset) never defined `.max-h-[90vh]` (nor
`.max-h-40/48/56/64/72/80`). The card's max-height computed to `none`, so it grew
to its full ~11000px content height. `.max-h-60` was the only max-h defined.

CSS-only fix in src/index.css — add the max-h-* scale (Tailwind N x 0.25rem) and
the arbitrary `.max-h-\[90vh\]`. Capping the card activates its already-correct
`flex-1 overflow-auto` body (overflow != visible zeroes the flex item's auto
min-size, so no min-h-0 needed). Repairs every consumer of these classes app-wide
(modal cards, dropdowns, scroll panels), not just this modal. No component JS
changed.

Verified live: card max-height resolves to 90vh (was none), header/footer both
in-viewport, body scrolls (scrollHeight > clientHeight). Extended
src/index.overlays.test.js (the #279 guard) with value assertions for each
max-h-* utility plus a check that no max-h-* class used in JSX is left undefined.
…scroll

Fix Assessments New-Assessment modal: define max-h-* so the card caps at 90vh
…-granularity

Scoring granularity: quarter-point second dropdown + 5-point CMMI scale option (#277)
…n from the org profile — no AI, no API key

The org-profile tailoring tier was name-substitution only: a profile
naming Google Cloud still produced procedures reading 'AWS CloudTrail,
GuardDuty, SentinelOne' — the checkbox looked like tailoring but changed
nothing about the environment. All real adaptation was delegated to the
optional AI tier.

This adds a curated substitution engine (stackTailorMaps.js) built from
the community bank's actual vocabulary:

- Cloud: AWS service references map to real analogs (CloudTrail → Cloud
  Audit Logging / Azure Activity Log, GuardDuty → Security Command
  Center / Microsoft Defender for Cloud, S3 → Cloud Storage / Blob
  Storage, ...). Fires only when a non-AWS cloud is selected without
  AWS. On-prem-only profiles get capability-language neutralization,
  never fake console steps.
- EDR: SentinelOne swaps to a named product (CrowdStrike Falcon,
  Defender for Endpoint, ...); a generic 'EDR' chip neutralizes to
  'your EDR platform'; never guesses a vendor.
- Email/chat: O365→Google Workspace and Slack→Teams/Google Chat on
  explicit signals only.

Matching is case-sensitive, word-bounded, longest-phrase-first — the
bare-AWS fallback can never fire inside 'AWS Security Hub', and
evidence paths (EVD-aws-*.md) pass through untouched. Applying twice is
a no-op.

Surfaces: second wizard checkbox with a live swap-plan disclosure (and
an honest explanation when no swaps apply), plus a per-item 'Adapt to
my environment' action so existing assessments benefit without
re-creation. Swapped output carries tailored provenance, so the
share-export pristine swap covers it — proven by a new producer-path
leak test. Full-bank sweeps assert zero AWS-vocabulary residue across
all 106 procedures for GCP, Azure, and on-prem targets.
…oring

Deterministic stack-aware tailoring: canned tool/platform substitution from the org profile (no AI)
…SF frameworks

Settings.js: replace ad-hoc Tailwind (bg-white, amber/yellow utility
classes) with the app's card/callout/settings-section-* primitives so
the page reads consistent with the rest of the terminal-style chrome.

index.css: add the layout utilities and callout/section-title styles
Settings.js now depends on (mx-auto, space-y-8, right-2, callout
variants) — these were referenced but never defined.

frameworksStore.js: bump persisted store to v6 and drop the pre-seeded
ISO 27001 / FedRAMP / CMMC stub frameworks (never fully implemented,
comingSoon flags never resolved). The app is NIST CSF 2.0 only now;
users bring any other framework via Import New Framework. Migration
filters retired IDs out of existing persisted state and preserves any
user-imported custom frameworks.
…cademy-citation

Cite Simply Cyber Academy for the 5-point CMMI scoring legend
…ls (#284)

Assessment setup (wizard step 1) gains an External Tracking option: declare
that findings, artifacts, and controls live in your own ticketing/compliance
system and name it (Jira, ServiceNow, SharePoint, ...). The name personalizes
the link-field labels; the URL fields themselves are always available so no
assessment is ever locked out of linking.

- findings.externalUrl (new) — ticket link, shown in the detail panel and as
  a link-out icon in the list
- controls.externalUrl (new) — this control in your compliance/ticketing tool
- artifacts.link (existing) — now labeled for ticket / Google Drive /
  SharePoint use
- assessments.externalTracking {enabled, systemName} — stamped by a v11
  store migration (ASSESSMENTS_SCHEMA_VERSION 10 -> 11); restore of older
  backups stamps the disabled default

Security:
- sanitizeExternalUrl: every user-pasted URL renders as an anchor only if it
  parses as absolute http/https; javascript:/data:/protocol-relative render
  as plain text. Applied to ALL anchor sites including the pre-existing
  artifact link and framework sourceUrl anchors (latent javascript:-href
  vector closed).
- Share export scrubs findings.externalUrl, artifacts.link,
  controls.externalUrl, and the system name by default (internal hostnames /
  project keys are private data); include-private opt-in keeps them.
  NOTE: artifacts.link previously rode share exports untouched — this is a
  deliberate behavior change, documented in PRIVATE_DATA.md.
- Scrub is copy-on-write; live store records are never mutated (tested).

Tests: 363/363 (37 new/updated: sanitizer unit tests, v11 migration,
restore stamping, whole-envelope canary leak tests both polarities).
Also adds five missing index.css utilities (hover:underline, inline-flex,
break-all, gap-1.5, ml-1) that existing pages already referenced.
…al-ticketing

External ticketing/document links for findings, artifacts, and controls (#284)
Discontinue the live Jira/Confluence integration and park it under
_future/atlassian-integration/ for reference. Removes from the app:

- Settings "Atlassian API Configuration" card + Confluence entry-ID harvesting
- src/utils/confluenceSync.js and Atlassian env validation
- Atlassian-specific error UI in ErrorBoundary
- Smart-embed usage in the assessments CSV export
- Server /api/jira, /api/confluence, /api/config routes/controllers/services
- SET_UP_IN_ATLASSIAN/ docs and atlassian-integration/ scripts
- Atlassian env vars, tauri CSP hosts, README/CONTRIBUTING/docs references

Kept: CSV import/export (Jira/Confluence-compatible format), the jiraKey
record field, and the Claude AI Assistant. Build clean; 24 suites / 353 tests pass.
@pull pull Bot locked and limited conversation to collaborators Jul 19, 2026
@pull pull Bot added the ⤵️ pull label Jul 19, 2026
@pull
pull Bot merged commit 64e917f into ericrihm:main Jul 19, 2026
5 of 6 checks passed
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant

0