Skip to content

mmaideveloper/aiarchitecttoolkit / skills/ai-solution-final-review

ai-solution-final-review

Verified

Perform a final, evidence-backed readiness review of an AI solution by synthesizing architecture conformance, AI behavior and evaluation, security, platform, operations, governance, and unresolved risks. Use before an AI solution release or formal decision gate; do not use as the sole specialist security, legal, or cloud review.

FileKindSize
SKILL.mdskill md3.6 KB
agents/openai.yamlconfig239 B
references/final-review-method.mdmarkdown1.4 KB

references/final-review-method.md

View on GitHub
# AI Solution Final Review Method

Use this ledger:

| Gate area | Requirement or risk | Required evidence | Current evidence and date | Status | Owner/action |
|---|---|---|---|---|---|
| <area> | <traceable requirement/risk> | <test, review, approval, runbook, telemetry> | <source and scope> | Pass / Conditional / Fail / Not verified / Not applicable | <owner and action> |

Minimum gate areas when applicable:

- Product acceptance and architecture conformance
- Model/provider/version identity and change controls
- Representative offline and online evaluation, thresholds, datasets, limitations, and regression evidence
- Threat model, security review, privacy/data lifecycle, safety, abuse controls, and human oversight
- Reliability targets, load/capacity, dependency failure, recovery, rollback, observability, incident response, and support ownership
- Transparency, user recourse, accessibility, records/traceability, third-party terms, and regulatory applicability
- Cost envelope and operational sustainability

Evidence is current only when it matches the reviewed release candidate and environment, remains within its validity period, and covers the relevant users, languages, data, failure modes, and integrations. Note sampling limits and uncertainty. A condition must be specific, measurable, owned, time-bound or trigger-bound, and non-blocking under an evidenced gate policy.
rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 194kUpdated Claude CodeCodexVerifiedInstall