Musah Abdulai
I take AI prototypes to production. ClearWIP founder.
- 1
- Followers
LLM Pre-Flight turns the last mile of an AI launch into a visible, shared readiness process. It helps teams document evidence and decisions around prompt injection, PII leakage, prompt extraction, access controls, observability and runaway cost.
I designed and shipped the...
What a 48-hour Controls Review looks like when it lands: findings with severity, evidence, fix path, owner and effort across IAM, logging, CI/CD, vulnerability, backup and AI workload controls. Demo data for a fictional company; the format and depth are the real deliverable.
34 free, MIT-licensed templates for SOC 2 and customer security review evidence: access controls (MFA, IAM, offboarding), logging and retention, CI/CD, vulnerability management. Covers AWS, GCP, Azure and GitHub. Turns 'trust me' into artifacts an auditor or enterprise buyer can verify.
A production-ready RAG starter where guardrails are part of the architecture: an input guard that blocks prompt injection and jailbreaks, an output guard that redacts PII (emails, SSNs, cards, API keys), and rate limiting. Qdrant semantic search, FastAPI, Next.js, PostgreSQL, plus a Terraform deploy repo.
A local CLI that tests LLM and RAG apps before launch for the failures that hurt in production: prompt injection, system-prompt extraction, PII leakage, refusal bypass, hallucination, cost. 7 attack modules, HTML audit reports with evidence and remediation, extensible pack...
