LLM Vault
Who We ServeThe Control ModelRoadmapLessons LearnedBlogAbout UsSign InBeta Now
Lessons learned/Tags

Every Tag, As A Permalink.

The unit of the registry is the failure, not the report. Every tag below is addressable — click through to see every failure across every report that carries it, no matter how many reports the corpus grows to.

Failure class
#constraint-ignored1#over-commitment1#self-cancelling-instructions1#no-unit-cost1#empty-output-counted1#false-liveness1#undeliverable-format1#silent-scope-reduction1#placeholder-shipped1#single-source-join1#default-as-measurement1#loose-pattern1#negation-blindness1#over-claiming1#environment-cause1#validator-conflict1#wrong-artefact1#correction-not-generalised1#unreconciled-figure1#self-report-trusted1
Layer
#planning1#prompting1#orchestration1#execution1#data-join1#verification1#reporting1#delivery1#environment1#measurement1
Cost driver
#token-waste1#rework1#abandoned-work1#redundant-retrieval1#human-time1
Detection
#caught-by-user1#caught-by-guard1#caught-by-luck1#caught-late1#silent1
Outcome
#near-miss1#shipped1#unrecovered1
LLM Vault

The governed layer between your organisation and every model, agent, router and coding tool it uses.

LLM Vault is tooling and evidence to help you manage AI risk — it does not certify compliance.

Explore

  • Who We Serve
  • The Control Model
  • The Governance Loop
  • Roadmap

Company

  • About Us
  • Blog
  • Contact

Trust

  • Privacy
  • Data Usage
  • Security

© 2026 CyberEQ · LLM Vault. All Rights Reserved.

Founded By John Mackenzie · AustraliaAdmin

DashboardReportsTagsFeed