Implement Stage A: rule engine, validation, audit, and safety gates

Completes 109 of 121 tasks. Every remaining task needs a tenant connection
(T055, T056, T101-T103) or an Azure Automation account (T115-T121).

  354 offline Pester tests      PASS
  Engine purity (Principle IV)  PASS
  Sanitization (SC-013)         PASS  (156 files)
  Graph module loaded in tests  none  (SC-008 holds)

What landed
  - Four-layer configuration validation with stable finding codes, covering
    every VR-002 and VR-003 condition, plus a 23-fixture invalid-config corpus
  - Run loop, audit records (NDJSON through a single sink), summaries,
    reconciliation, and exit codes 0-6
  - Persistence behind a single write-body builder whose result always has
    exactly one key
  - Invoke-PersonaEngine.ps1 and Edit-PersonaEngineConfig.ps1
  - Six docs, two pipelines, traceability matrix, V-5a and sanitization records

Three deviations from tasks.md, each recorded in its status block

  T033 is not in Resolve-UserPersona. evaluationErrorThreshold is run-level
  state and the rule engine is pure; a counter there would break Principle IV.
  It lives in New-PersonaRunCounter and is applied in the run loop.

  A new src/Engine/ layer holds Invoke-PersonaEngineRun. The entry script
  imports the manifest, which requires Microsoft.Graph.Authentication, so a
  loop living only inside it could not run on a machine without the Graph SDK
  and SC-004 could not be proven at all. The entry script is now a thin
  wrapper and what ships is what is tested.

  The invalid-config corpus is generated by a committed script, with the
  generated fixtures committed too, so a reviewer sees the fixture in the diff.

Defects found by running the code, not by reading it

  Group and role ID lists were double-wrapped: @(Get-PersonaGroupIdPage ...)
  around a comma-returned array collapsed every membership list into one
  bogus space-joined entry. That is a silent false non-match, exactly what
  FR-013 exists to prevent.

  A 403 whose status appears only in the exception message parsed as $null,
  which the retry policy treats as a transport error - five requests per
  account against a tenant already refusing. Status extraction now falls back
  to the message text, bounded to 400-599.

  The sanitization scan walked tracked files only, so it covered 34 of 156
  files and none of this phase's code. It now scans untracked non-ignored
  files too, and a negative control confirms it catches a planted leak.

  Test-Json reports one error per violating location, not first-failure-only
  as the V-5a draft claimed. Record and pin corrected.

Enforcement remains blocked on the V-4 security sign-off.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This commit is contained in:
2026-08-20 21:48:19 -04:00
parent c59c85dd55
commit cdc6bb33d3
124 changed files with 16638 additions and 199 deletions
@@ -0,0 +1,127 @@
# Contract: Structured Audit Records
Serialized form of the audit trail (FR-022, NFR-005, Principle V). Format is newline-delimited JSON,
one record per line (OTD-006). All emission goes through a single `Write-PersonaAuditRecord` sink so
a future transport can be added without touching call sites.
## Common envelope
Every record carries:
| Field | Type | Notes |
| --- | --- | --- |
| `timestamp` | string (ISO 8601 UTC) | |
| `recordType` | string | `RunStart`, `UserEvent`, `Summary`, `RunComplete`, `EngineDefect` |
| `runId` | string (GUID) | Constant for the run (NFR-005) |
| `engineVersion` | string | |
| `configVersion` | string | |
| `configurationHash` | string | SHA-256 of the configuration file |
| `mode` | string | `Preview` or `Enforce` |
## `UserEvent`
Emitted once per processed user. 100% of these records carry `runId`, `userPrincipalName`, and
`accountObjectId` (SC-006).
```json
{
"timestamp": "2026-08-20T09:14:02.187Z",
"recordType": "UserEvent",
"runId": "<RUN-ID>",
"engineVersion": "1.0.0",
"configVersion": "1.4.0",
"configurationHash": "<SHA256>",
"mode": "Preview",
"accountObjectId": "<ACCOUNT-OBJECT-ID>",
"userPrincipalName": "<USER>@<PRIMARY-DOMAIN>",
"outcome": "Matched",
"matchedRuleId": "RULE-0100-TIER0",
"storedPersona": "Employee",
"calculatedPersona": "Tier0-Admin",
"previousValue": "Employee",
"action": "WouldUpdate",
"rulesEvaluated": 4,
"durationMs": 38,
"evaluationErrorReason": null
}
```
**Field requirements**
| Field | Requirement |
| --- | --- |
| `outcome` | Exactly one of `Matched`, `Unclassified`, `EvaluationError` (SC-001) |
| `matchedRuleId` | Non-null on every `Matched` record (SC-006) |
| `previousValue` | **Captured at write time on every `Updated` record.** This is what makes OTD-010 rollback possible; omitting it in v1 makes rollback impossible retroactively |
| `evaluationErrorReason` | Non-null exactly when `outcome` is `EvaluationError` |
**Prohibited fields**: access tokens, `Authorization` headers, secrets, and full Graph responses
MUST NEVER appear in any record.
**Condition tracing**: a `conditionTrace` array may be added **only** under `-Debug`
(`logging.traceConditionValues`). It contains diagnostic condition-level values and is therefore
gated by acknowledgement (VR-003).
## `Summary`
Emitted every `summaryInterval` users and once at completion.
```json
{
"recordType": "Summary",
"runId": "<RUN-ID>",
"summaryType": "Interim",
"processed": 250,
"matched": 231,
"unclassified": 14,
"evaluationError": 5,
"unchanged": 220,
"wouldUpdate": 11,
"updated": 0,
"updateFailed": 0,
"reconciliationPassed": true,
"ruleCounts": [
{ "ruleId": "RULE-0100-TIER0", "name": "Tier 0 administrators", "enabled": true, "matches": 3 }
]
}
```
`reconciliationPassed` is `processed == matched + unclassified + evaluationError` (FR-021, SC-007).
`ruleCounts` lists **all** business rules, including disabled and zero-match rules — an absent rule
is indistinguishable from a rule that never fired, and operators need that distinction.
## `RunComplete`
```json
{
"recordType": "RunComplete",
"runId": "<RUN-ID>",
"startedUtc": "2026-08-20T09:00:00.000Z",
"completedUtc": "2026-08-20T09:12:44.913Z",
"durationMs": 764913,
"processed": 4820,
"matched": 4611,
"unclassified": 190,
"evaluationError": 19,
"unchanged": 4400,
"wouldUpdate": 211,
"updated": 0,
"updateFailed": 0,
"reconciliationPassed": true,
"exitCode": 0
}
```
## `EngineDefect`
Emitted when reconciliation fails (FR-021) or an internal invariant is violated. Severity is always
`Error`. A failed reconciliation is a defect in the engine, not a property of the data, and is
reported as such rather than being folded into ordinary counters.
## Sanitization (SC-013)
Committed artifacts — this contract, examples, fixtures, tests, and documentation — use placeholders
only: `<ORGANIZATION-NAME>`, `<PRIMARY-DOMAIN>`, `<TENANT-ID>`, `<ACCOUNT-OBJECT-ID>`,
`<GROUP-OBJECT-ID>`, `<APPROVED-PERSONA-ATTRIBUTE-NAME>`, `<AUTOMATION-ACCOUNT-NAME>`,
`<LOG-OUTPUT-PATH>`, `<RUN-ID>`. Runtime records naturally contain real UPNs and Object IDs — which
are approved for logs — but no such value may ever be committed to this repository.