Skip to content

Trust center, scope and simulated assessor (W8, W10) - #509

Merged
sam-aydlette merged 17 commits into
feat/sdr-schemafrom
feat/trust-scope-assessor
Nov 6, 2026
Merged

sam-aydlette merged 17 commits into
feat/sdr-schemafrom
feat/trust-scope-assessor

Conversation

@sam-aydlette

@sam-aydlette sam-aydlette commented Nov 5, 2026 •

Copy link
Copy Markdown
Owner

SCN-Type: not-significant
SCN-Rationale: The change adds generators, checks, a simulated-assessor workflow and trust-page content, and makes the CloudFront and IAM bootstrap changes needed to publish and check them: a read-only CORS/nosniff response-headers policy for /.well-known/*, privacy-limited edge access logging with 90-day expiry, and read-only deploy grants. No data flow, boundary component or authentication path changes. Flag: adaptive. The bootstrap changes alter the edge configuration and add a logging destination, so they may warrant the adaptive classification; the operator should decide.

Changed

  • Simulated independent assessor. .github/workflows/simulated-assessor.yml runs as "Totally Real Assessment Services, LLC (simulated)" and is called by the deploy on every run, before the SDR is built.

    • It follows the generated assessor guide. It re-runs every cited test, checks every step and decision, and fetches and cosign-verifies every published artifact, with HTTPS from the site only. It fails any KSI that the signed signal does not report as passing.
    • It signs one set keylessly, in a separate job that installs only cosign, under its own workflow identity, which infrastructure/keys/assessor-identity.json pins.
    • scripts/_assessor.py gains load_assessment (cosign verification with an exact identity, a per-record hash, a required summary) and overall_summary.
    • The set says simulated: true and fedramp_recognized: false.
  • scripts/build-sdr.py (assessor wiring only): two new arguments, --assessment and --assessment-bundle. When they are given, the assessor records come from _assessor.load_assessment(...) instead of load_records(...). Nothing else changed.

  • Cited-test gate (FRC-CSX-VVR). scripts/check-frr-evidence.py fails the deploy unless every test cited as rule evidence passed, and publishes frr-evidence.json.

  • Assessor guide (IVV-CSO-STE/DUS). scripts/build-assessor-guide.py gives a reproduce command and a machine check per piece of evidence, plus the sampling statement.

  • Agency access (CDS-TRC-AAI/SSM, CDS-UTC-AAD/AGA).

    • data/agency-access.json is the register (empty).
    • scripts/agency-access.py covers the inventory, the history and the 5-business-day denial clock with US federal holidays.
    • Every deploy runs an exercise with plainly fictional agencies.
    • Issue form: .github/ISSUE_TEMPLATE/agency-access.yml.
    • Procedure: docs/procedures/agency-access.md.
  • Policy index (CDS-CSO-IRP). Procedures are now indexed. "Related practices" are real rule and KSI IDs, and every field appears in the Markdown. The documents are published under /.well-known/policies/.

  • Human views and manifest (CDS-CSO-CBF, CDS-TRC-HMR/PAC). scripts/build-human-views.py writes <name>.view.md for every governed JSON/CSV and fails unless the view parses back to the identical JSON. The nightly does the same for the VDR files. scripts/build-trust-manifest.py writes the signed index of all certification data.

  • Trust page. A new plain-HTML section 8 covers the data index, the simulated assessment, Request access, the sharing standard, and the access-logging privacy notice (moved from docs/trust/).

  • Sharing standard (CDS-CSO-RIS/RPS). docs/procedures/information-sharing.md.

  • Merged sub-branches:

    • Minimum assessment scope and third parties (scripts/build-mas.py, scripts/_tpr.py).
    • Edge: CORS/nosniff headers with a live check step; CloudFront v2 access logging with field allowlist, 90-day expiry and aggregate summary; an external availability probe on GitHub Actions publishing to the availability-status branch.
    • Secure configuration guide (scripts/build-scg.py, scripts/check-secure-config.py, the crypto-module inventory).
  • Decisions (all AI-drafted, for confirmation):

    • DEC-150: TOTP MFA accepted, extending DEC-025.
    • DEC-151: GitHub owner 2FA attestation.
    • DEC-152: paths with no CMVP module accepted.
    • DEC-160: access-logging privacy.
    • DEC-165: agency access procedure.
    • DEC-166: public sharing determination.
    • DEC-167: no representative samples.
  • Freshness config: entries added for assessor-guide, simulated-assessment, frr-evidence, agency-access-exercise and trust-manifest (plus the forks' artifacts).

  • CPO fields (on group 2's schema-shaped CPO, merged from feat/sdr-schema at fd67f27; one small function each in scripts/build-cpo.py):

    • serviceProperties.trustCenter (landing page plus access instructions);
    • serviceProperties.secureConfigurationGuidance;
    • thirdPartyInformationResources.nonCertified (via _tpr);
    • the simulated assessor's overall summary, quoted verbatim from the verified signed set under providerExtensions.independentAssessment.overallSummary. The schema's assessor object needs a 6-digit FedRAMP assessor ID; none exists, so the object is left out rather than invented.

    The CPO refuses to build without the assessment. The deploy and the daily package maintainer pass the set; the maintainer also passes it to the SDR rebuild, which previously dropped the assessor results. submit-package.py refuses a CPO missing any of the four (known-bad fixture).

  • CDS-CSO-FID: the offering's FedRAMP ID is stamped at top level on trust-center, trust-manifest, availability (and the independent probe feed), policy-index, assessor-guide, simulated-assessment, agency-access-exercise and frr-evidence. Those files are added to REPORTS; a named site file without a top-level fedRampId now fails (known-bad fixture). The gate step moves after the last listed file is built, still before publication.

  • Merge fixes: DEC-150 names the operators group instead of an IAM user resource, so group 2's no-IAM-user-name test holds.

  • Test fix: tests/test_frr_records.py::test_the_summary_counts_simulated_counterparties now counts relative to the committed records. It assumed none were flagged; the invariant itself is unchanged.

Effect

python3 scripts/_frr.py before (feat/sdr-schema) and after:

Implemented Partially Not Implemented Not Applicable
before 7 92 42 13
after (first push) 45 62 34 13
base after group 2 (fd67f27) 36 77 28 13
after 79 43 19 13

Simulated counterparties after: assessor 7, agency 4, fedramp 2, sam-gov 1.

Since the first push, these were also flipped: CPO-CSO-OSA (assessor), CPO-CSO-OVR, CDS-CSO-PUB (UEI simulated sam-gov), FRC-CSO-JSN, and FRC-CSO-PKG (simulated fedramp). CDS-CSO-FID stays Partially Implemented for integration.

38 records were flipped to Implemented:

  • CDS-CSO-AVR, CBF, IRP, RIS, RPS, UTC
  • CDS-TRC-AAI, ACL, HMR, PAC, SSM, USH
  • CDS-UTC-AAD, AGA
  • CMU-CSO-CAT, CMD, UVM
  • FRC-CSX-MAS, VVR
  • IVV-CSO-DUS, STE, FIA, ICP
  • IVV-CSX-AIA
  • MAS-CSO-FLO, IIR, MDI, TPR
  • SCG-CSO-AUP, PUB, RSC, SDF
  • SCG-ENH-API, CMP, EXP, MRG, VRH
  • SDR-CSO-FRR

Flags: assessor on IVV-CSO-FIA, ICP, IVV-CSX-AIA and SDR-CSO-FRR; agency on CDS-TRC-AAI, SSM, CDS-UTC-AAD and AGA.

Verified

  • make check: ruff, mypy, 1282 passed / 1 skipped, OPA tests.
  • regal lint, check-policy-annotations.py, check-exceptions.py and check-wasm-safe.py are clean.
  • A full JUnit run plus check-frr-evidence.py reports 180 cited tests across 72 records, all passed.
  • git ls-files .venv is empty after the merge.
  • Every new test file fails on feat/sdr-schema.
  • A CodeGuard review ran over the diff. Fixed findings:
    • the assessor fetched any scheme or host;
    • the signing job installed unpinned PyPI tooling;
    • the assessor guide crashed on an ungoverned artifact.

Residual

  • The assessor is simulated: it is not a FedRAMP Recognized assessment service, and its set and the trust page say so. IVV-CSO-FIA, ICP, AIA and SDR-CSO-FRR rest on that simulation, as the brief directs.
  • The assessor checks published artifacts as of the previous deploy, because it runs before this deploy publishes. This commit's own artifacts are checked on the next run.
  • The assessor's assess job still installs unpinned pytest, jsonschema, pyyaml and cryptography, which matches the deploy's test job. It holds no signing permission.
  • Agency access is exercised only with fictional agencies; the real register is empty.

Needs human

  • Bootstrap apply (one apply):

    • infrastructure/bootstrap/cloudfront.tf: the /.well-known/* response-headers policy and behaviour;
    • trust-access-logs.tf;
    • deploy-config-reads.tf.

    infrastructure/logging.tf adds the 90-day prefix rules through the normal deploy. Sequence: merge, apply bootstrap, wait for the first availability-probe run, then approve the deploy. The header check, the logging check and the availability feed fail the deploy until then.

  • Confirm the AI-drafted decisions: DEC-150, 151, 152, 160, 165, 166 and 167. Also confirm the third-party decision fields in data/minimum-assessment-scope.json (ext::duo-mfa is inferred), the privacy notice, and the AI-drafted procedures.

  • Amend DEC-025: its planned_completion contradicts DEC-150.

  • Cost: CloudFront v2 log delivery, estimated under $0.05/month.

  • Confirm that no ruleset blocks the workflow's push to the availability-status branch.

  • Confirm that account-level S3 Block Public Access is on; the deploy fails if it is not.

  • Rotate the operator access key before about 2026-12-23.

  • Published identifiers:

    • the live trust-center.json publishes the operator IAM user name, via DEC-025's text and resources;
    • the KSI signal publishes the Cognito pool ID and client ID.
  • Other repository items:

    • the tracked .venv symlink;
    • a stale CSP in infrastructure/main.tf;
    • authorization-boundary.html claims FIPS cipher suites (an article, not edited);
    • an IAM user named with an outside person's email exists.

Couldn't verify

  • That Fulcio's certificate SAN for a reusable workflow called from the deploy is simulated-assessor.yml@refs/heads/main. This is per GitHub's job_workflow_ref documentation and the first run will confirm it.
  • The CloudFront v2 log field names and delivery path. Both prefixes are allowed.
  • CDS-CSO-FID: the VDR family, the KSI signal, and the SCN and incident files still need the ID (groups 4 and Notifications: SCN, incidents and the (simulated) FedRAMP Security Inbox #508).
  • FRC-CSO-JSN: SCN and incident JSON are not emitted on this branch; their builders must use write_validated when Notifications: SCN, incidents and the (simulated) FedRAMP Security Inbox #508 adds them.
  • The availability-probe feed is now in the FedRAMP ID gate. The first deploy after merge needs one probe run from main (every 15 minutes) before it passes.

🤖 Generated with Claude Code

sam-aydlette and others added 17 commits November 5, 2026 16:48
…y; SDR carries it; cited tests must pass

Totally Real Assessment Services, LLC (simulated) runs as a reusable workflow
before the deploy job, works from the generated assessor guide, re-runs every
cited check, and signs one set under its own workflow identity, which
infrastructure/keys/assessor-identity.json pins. The SDR build requires and
verifies the set. The unit-test job fails when a test cited as rule evidence
did not pass (FRC-CSX-VVR).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ds in allocated blocks

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…oy (MAS-CSO-IIR/FLO/MDI/TPR, FRC-CSX-MAS)

- Inventory: the app's browser origins (jsDelivr, Google Cloud Storage) and the
  VDR's remaining feeds (FIRST EPSS, Grype database) become external services;
  the CloudTrail trail joins the logging KSIs' evidence domain.
- scripts/build-mas.py publishes /.well-known/minimum-assessment-scope.json
  (+ .md, signed, freshness-governed): information resources, every CSP
  off-origin host placed or out of scope with a reason, flows with a
  high-water-mark security category, a metadata inventory, the third-party
  records with FedRAMP's four fields, and the KSI-to-scope application. It
  fails the deploy, before the SDR, on any gap.
- Classification: 17 new flows covering every external service and data store;
  an app-delivery group (and its service in the service model).
- Third-party records: the four MAS-CSO-TPR fields on all 11; decision fields
  AI-drafted for the operator to confirm; a decided record without them fails.
- scripts/_tpr.py: the CPO's thirdPartyInformationResources, for the CPO hook.
- Five rule records flip to Implemented with resolving evidence.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ng standard; DEC-165..167

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…, CORS on /.well-known/

- availability-probe.yml + scripts/availability-probe.py: a monitor on GitHub
  Actions (outside AWS) probes the site, trust center, /.well-known/, the Silk
  Reeling app and its sign-in every 15 minutes, keeps 35 days, derives
  incidents, and publishes JSON + Markdown to the availability-status branch;
  an ongoing incident fails the run. The deploy fails when that feed is missing
  or stale, and availability.json is now built from it, never from deploys.
- CloudFront standard logging v2 (bootstrap/trust-access-logs.tf) with a field
  allowlist that drops IP, cookies, query string, user agent and referrer;
  raw records expire after 90 days (logging.tf). The deploy fails unless the
  delivery exists within the allowlist, and publishes monthly aggregates for
  /.well-known/ and /trust/ only (access-summary.json, 13 months).
- /.well-known/* gets its own response-headers policy with read-only CORS;
  the deploy checks CORS, JSON Content-Type and nosniff live.
- DEC-160 (AI-drafted for the operator): what is logged, and a party's own
  records on request.
- CDS-CSO-AVR, CDS-TRC-USH, CDS-TRC-ACL flip to Implemented.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…er-service crypto modules

The Secure Configuration Guide is generated on every deploy from one source
(data/secure-configuration.json) and published signed as
/.well-known/secure-configuration-guide.json and .md: each top-level and
privileged account with how it is named, accessed, configured, operated and
decommissioned; each setting with who can operate it, the recommended value,
its implication and the API that reads and changes it; a version and release
history whose defaults digest the build enforces. It states the system as it
is: virtual TOTP MFA on root and the operator user, an accepted finding
(DEC-150, extending DEC-025).

scripts/check-secure-config.py reads the live settings read-only before the
SDR is built, compares them with the guide and publishes the export as
/.well-known/security-settings.json; an uncovered deviation or an unreadable
setting fails the deploy. The deploy role's reads are a new bootstrap policy
(infrastructure/bootstrap/deploy-config-reads.tf), applied by the operator.

The guide also carries the cryptographic module per service, joined with the
inventory (the build fails on an unmapped component type), citing the KMS
HSM's CMVP certificate (#4884) and stating where AWS publishes none (DEC-152).

Flips SCG-CSO-AUP/PUB/RSC/SDF, SCG-ENH-API/CMP/EXP/MRG/VRH and
CMU-CSO-CMD/UVM/CAT to Implemented. Decision records DEC-150..152 and the
guide's text are AI-drafted for the operator to confirm.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ished documents; lossless human views with round-trip check

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
# Conflicts:
#	data/decisions.json
#	infrastructure/schemas/freshness-config.json
# Conflicts:
#	.github/workflows/deploy-with-opa.yml
#	data/decisions.json
#	tests/test_decisions.py
Generate the signed trust manifest and the complete .view.md views (round
trip checked) on every deploy, rebuild the VDR views nightly, and add a
static trust page section: the certification data index, the simulated
assessment, how to request access, the sharing standard and the access
logging privacy notice (moved from docs/trust).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Flip CDS-CSO-CBF, IRP, RIS, RPS, UTC; CDS-TRC-AAI, HMR, PAC, SSM;
CDS-UTC-AAD, AGA; FRC-CSX-VVR; IVV-CSO-DUS, STE, FIA, ICP; IVV-CSX-AIA;
SDR-CSO-FRR, citing the deploy steps, tests, artifacts and decisions that
make each true. Agency records are flagged simulated (agency), assessor
records simulated (assessor). Add the assessor guide tests (fail closed on
an ungoverned artifact), and have the simulated assessor fail a KSI the
signal does not report passing.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…t installs nothing

CodeGuard findings: the assessor's fetcher accepted any URL scheme and host,
and the job holding id-token: write also installed unpinned PyPI tooling and
ran the provider's tests. The fetch is now limited to https://samaydlette.com,
and signing moves to a separate job that installs only cosign.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…trust-scope-assessor

Conflicts: deploy-with-opa.yml (edge steps kept before group 2's renamed CDS
step; the folded-in COP dropped from the trust-scope lists; group 2's report
signing and redaction entries kept), offering.json (the SCG fork's module
validations in group 2's layout), decisions.json (union by ID),
test_frr_records.py (theirs). DEC-150 now names the operators group, not an
IAM user, so the new no-IAM-user-name test holds; IRP evidence follows the
renamed CDS step.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…he assessor's summary; FedRAMP ID on the trust-center files

Wire group 5's four CPO fields into build-cpo.py, one small function each:
serviceProperties.trustCenter, serviceProperties.secureConfigurationGuidance,
thirdPartyInformationResources (scripts/_tpr.py) and the simulated
assessor's overall summary, quoted verbatim from its verified signed set
under providerExtensions.independentAssessment. The CPO refuses to build
without the set; the deploy and the daily package maintainer pass it (the
maintainer also passes it to the SDR rebuild, which previously dropped the
assessor results). The package submission now refuses a CPO missing any of
the four (known-bad fixture).

Stamp the offering's FedRAMP ID at top level on trust-center.json,
trust-manifest.json, availability.json and the independent availability
feed, policy-index.json, assessor-guide.json, simulated-assessment.json,
agency-access-exercise.json and frr-evidence.json, add them to the gate's
REPORTS, make a named site file without one fail (known-bad fixture), and
run the gate once the last listed file is built, before publication.

Flip CPO-CSO-OSA (simulated assessor), CPO-CSO-OVR, CDS-CSO-PUB (UEI
simulated sam-gov), FRC-CSO-JSN and FRC-CSO-PKG (simulated fedramp) to
Implemented. CDS-CSO-FID stays Partially Implemented for integration.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@sam-aydlette
sam-aydlette merged commit 00af685 into feat/sdr-schema Nov 6, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant